Men who design weapons for a living will gather in Huntington Beach, Calif., Tuesday to learn about the Pentagon's latest idea for stopping Soviet missiles in mid-flight.

The meeting under Army auspices at the McDonnell Douglas plant in Huntington Beach dramatizes that the search for the bullet that can stop another bullet continues, even though earlier approaches were found wanting.

First, there was Nike-Zeus: then Sentinel, then Safeguard - all names of ABM (anti-ballistic-missle) systems designed to stop Soviet missiles on the fly.

Each time the offense proved too much for the defence. Finally, the United States and Soviet Union in 1972 agreed to forgo deploying an extensive missile defense, although each side was allowed to keep trying to perfect one.

Since then both superpowers have been relying on their devastating retaliatory offense to discourage a first strike by the other.

However, Pentagon specialists assert, Soviet leaders have remained uneasy about relying solely on offense. Continuous experimentation with the Moscow ABM site after the 1972 treaty was signed and heavy spending on missile defense research manifest the Soviet uneasiness, according to the Pentagon specialists.

In contrast, U.S. leaders have shut down their only working ABM system - located at Grand Forks. N.D. - and have concentrated research money on defense systems that could not be put into action for years if the Soviets suddenly canceled the ABM treaty.

Even so, U.S. missile defense planners believe they are still ahead of the Soviet Union in the most promising technology. This technology will be described Tuesday to aerospace executives with security clearances.

Called the "overlay, underlay" defense backers claim it could keep land-based missiles from becoming obsolete at least through the 1980s. Concern that the U.S. force of 1,000 Minuteman missiles now standing in underground silos could not survive an attack by improved Soviet missiles is building up pressure for deploying the MX, a new missile that would be carted back and forth in a tunnel to make it hard to hit or destroy.

Deploying the MX, in the opinion of many arms controllers, would make the world much more dangerous. The missile would be so accurate and so powerful, they argue, that the Soviets in a crisis would be tempted to fire off their land-based missiles for fear MX would wipe them off. The Minuteman, smaller than the MX, is not viewed as a first-strike threat.

Another theory is that missiles are becoming too vulnerable if fixed on land and thus should be moved into bombers and submarines, abandoning one leg of the "triad" offense.

The layer defense would protect land-based missiles with two sets of rockets - one that would intercept incoming warheads in outer space and another that would attack them close to earth.

Unlike the Spartan anti-ballistic-missile which was built as part of the abandoned Safeguard defense, the "overlay" missile would not carry a nuclear warhead. It would be a "smart" missile that would fly into space and home in on the heat of the incoming warhead. The overlay missile would destroy the warhead by colliding with it or throwing rods or peliets tin its path.

The homing devices to go on the experimental missile were not available when Safeguard was deployed, according to Pentagon officials. The "overlay" missile would be so smart that it could not be fooled by decoy warheads released in space, a flaw in Safeguard.

Another advantage of overlay. Pentagon officials contend, is that the air defense commander probably would be allowed to fire ti since the warhead would not be nuclear. The President currently is supposed to be the only one to order a nuclear weapon fired.

Because the "smart" missile could hit the enemy warhead farthur out in space than any other ABM tested to date, it could cover more sky from one base than its predecessors. This would give the President more time to decide what to do in response to an attack, defense officials said.

For the moment, the plan is to design the close-in layer of missiles wtih nuclear warheads, largely because homing devices do not work as well in the atmosphere as in space. The close-in missile might be non-nuclear eventually.

The Pentagon is calling the longrange defensive missile HOE for homeing overlay experiment. The HOE and close-in missile, according to Army missile defense officials, could protect a Minuteman site for half the cost of the scrapped Safeguard defenses.

The giant, easy-to-destroy Safeguard radars were a major weakness of that defense. Small radars spread widely around the base are envisioned for the larger defense.

Defense officials said the HOE missile would be made from other rockets already on hand. The first HOE would not be launched until 1982 or 1983 under current Army plans.

Defense officials stress that the layer defense is still in the experimental stage under a $200 million annual missile defense research program. McDonnell Douglas executives at Tuesday's meeting will brief other firms on designs needed to advance the layer defense toward the flight test stage.

William A. Davis, deputy project manager of the Army ballistic misssile defense command in Huntsville, Ala., which is running this newwest ABM effort, said defending inercontinental ballistic missiles, still looks like a viable option "for assuring their survivability."