With a presidential decision nearing on the future of the Air Force's MX missile, a major battle over where and how to base the weapon is escalating in the Pentagon.

The Air Force, backed by some influential members of Congress, is trying to discourage the Reagan administration -- particularly Defense Secretary Caspar W. Weinberger-- from pressing ahead with a scheme to put the new MX missiles on transport planes.

Some senior Air Force officers have said privately they do not think they could testify in good conscience on behalf of that plan in Congress. Repeatedly in recent years many senior Air Force officers, including-chief of staff Gen. Lew Allen; deputy chief for research and development Lt. Gen. Kelly H. Burke and MX program directors in Washington and in Air Force ballistic missile commands, have testified that a land-based system is preferable to an airborne one. Allen has called an airborne system "feasible," but risky and costly.

The Air Force continues to prefer the original basing scheme, or possibly a scaled-down version of it. This is for a land-based system in which 200 missiles would be hidden and shuttled among 4,600 underground shelters in Utah and Nevada so the Soviets would not know where to attack to knock them out.

In the Joint Chiefs of Staff, senior military officers say leaders of other services are also concerned about the air-launched MX, fearing it would ultimately cost so much more than the land-based system that it would cut into money available for general purpose land, sea and air forces.

Weinberger has repeatedly said no MX decision has been made. But if the air-launched version is ordered, as he is reportedly ready to recommend, the Air Force will be in a position of fielding a system that much of its current hierarchy does not want as its primary nuclear retaliatory force for the future.

The private indications from some officers that they would now find it embarrassing to defend an airborne MX after years of rejecting it as a good solution, have raised speculation about resignations over the issue by some generals, though the officers themselves have not threatened this.

"This is not an emotional reaction," one senior officer said, "nor an attempt to pressure the administration. It's a professional question . . . of defending a fundamentally unsound thing. They simply will have to find somebody else to carry the mail" in defending the airborne plan, he said.

Should the administration go ahead with the airborne plan, there could be a replay of sorts of the long-running battle of the early 1960s when then Secretary of Defense Robert S. McNamara pressured the Navy and Air Force into using the same multipurpose TFX fighter-bomber-reconnassiance plane and then overruled a military selection board on which contractor should build the new aircraft.

An air-launched MX would ease some political problems for the president by removing the need for an enormous and controversial MX construction project in the West. Even before the election, Reagan was opposed to the western basing plan. It also might have some military advantages.

With even a small amount of warning time, some specialists claim that the missile planes could be in the air and thus have a better chance than land-based missiles of not being knocked out in a first strike by Soviet missiles. Each missile-transport combination would also immediately bolster the U.S. arsenal when delivered as opposed to the shelter-shuttle scheme, in which the entire project would have to be finished to confuse Soviet targeters and achieve security of individual missiles.

For a president whose party campaigned on the idea that this country faced a "window of vulnerability" to Soviet missile attack in the next several years, an air-launched MX, combined with a decision to build a modified version of the B1 bomber, would suggest that he is acting to close that "window" as fast as possible. The full MX land-based shelter system would not be available until the end of the decade, and a force of more advanced "Stealth" bombers would take even longer. A sizable number of MX missiles reportedly could be ready for action in C5 transports by 1986.

The air-launched MX also would not require an anti-ballistic missile (ABM) system to defend it. The United States therefore could avoid breaking the 1972 ABM treaty with Moscow, which would make it easier eventually to resume arms control talks with the Soviets if the administration chooses to do so. Critics of the shelter plan favored by the Air Force claim that it plays directly into Soviet strengths, meaning that the Russians could keep adding warheads to their big missiles to keep pace with an expanding shelter program.

Yet, outside of Weinberger's office, the airborne plan seems to have little support. An air-launched MX, critics argue, would have the same vulnerabilities bombers have, meaning they must get off the ground to survive an attack and then find a place to land. They might not be able to fly through a barrage of Soviet atomic bombs set off in the atmosphere above air fields. The missiles may be less accurate when launched from the air than from fixed silos and the planes may be harder for civilian and military commanders to communicate with and thus control.

There might be some early cost savings for the administration for an air-launched version in comparison to the Utah-Nevada scheme. But ultimately many experts believe it would be far more costly and require a large increase in manpower and airfield network to support it.

Under Weinberger's plan, some 115 new C5 transports would be used to carry MX missiles on an interim basis until a more advanced plane, nicknamed "Big Bird," is built. Then the C5s could be turned into troop transports to bolster Army airlift capabilities, a prospect that reportedly appeals to Weinberger.

Though the Air Force contends that the original 4,600-shelters plan is the best way to ride out an attack and respond with precision, that scheme seems dead, unless some extraordinary change of mind takes place. However, the administration could decide, perhaps in conjunction with an airborne system, to approve some kind of scaled-down land-basing for perhaps 100 MX missiles and silos in an area of Nevada where opposition is not strong. Such a possibility was reportedly among those suggested by the special 15-member Townes commission appointed by Weinberger last March to study MX basing alternatives.

Then, however, the question of missile vulnerability would immediately arise. If the current 1,000 U.S. land-based Minuteman missiles are vulnerable to attack by accurate Soviet land-based ICBMs, why wouldn't the new MX be just as vulnerable in fixed silos, or even in a reduced shelter scheme which the Soviets could overwhelm by adding more warheads?

The likely answer: ABM defenses could be built if necessary to protect the land-based MX. But a number of technical specialists in and out of the Pentagon argue that while the United States probably has a lead over Moscow in ABM technology and an ABM might complicate a Soviet attack, the offense continues to have the edge and missile defense basically won't work.

Also, to deploy an extensive ABM defense would require tens of billions of dollars more on top of already planned military projects. It would also mean tearing up the treaty with Moscow limiting such weapons. Both factors would certainly produce public debates which could wind up leaving the United States in the worst of all possible positions: with allegedly vulnerable missiles in land bases and no ABM to defend them.

The administration, therefore, faces a situation where there do not seem to be any "good" choices.

What drives the $35 billion to $100 billion MX program is the official claim that U.S. Minuteman missiles are vulnerable to Soviet attack. In a technical and theoretical sense, there seems agreement on this point by many specialists. Yet the idea that any Soviet leader would bet he could get away with even a limited attack on Minuteman, in which hundreds of atomic warheads would explode in the north-central United States, without massive retaliation from surviving missile-firing submarines and bombers is on the extreme edge of plausibility. Furthermore, neither country has tested its missiles over the course they would have to fly in wartime.

The Townes panel looked into the vulnerability question, according to Dr. Theodore Gold, the panel's executive secretary. Gold won't say what the panel found, though he says it did have access to secret data on alleged Soviet missile accuracy.

One of the studies Gold asked for was by J. Edward Anderson, a professor of mechanical engineering at the University of Minnesota who is a specialist in inertial guidance systems, the kind used in ocean-spanning missiles. Anderson, in his study and in a telephone interview, argued that a Soviet attack on U.S. missile bases cannot be expected to succeed for a variety of factors, including inherent guidance inaccuracies.