Retired Army Lt. Gen. Daniel O. Graham was identified incorrectly yesterday as a former Air Force officer.

A new study, sponsored by the conservative Heritage Foundation, proposes a major shift in U.S. defense strategy in which nonnuclear weapons shot from satellites in space and ground bases in this country would destroy Soviet missiles as they are flying toward the United States.

The study, which has been submitted to the Reagan administration, envisions a crash program to make "a technological end run" around the growing Soviet missile threat that would "nullify or substantially reduce" it.

It would take advantage of a U.S. edge in space technology, use equipment already in development, not require new American nuclear weapons and is not based on attacking missile silos in the Soviet Union.

In outlining the "High Frontier" study to reporters yesterday, retired Air Force Lt. Gen. Daniel O. Graham claimed it leads away "from the bankrupt and basically immoral precepts" of the strategy of mutual assured destruction in which each superpower threatens to blow up the other in retaliation for a first strike.

Pentagon officials who have seen the study say it ranges from "some very credible technology to complete Star Wars stuff." They view it as most unrealistic in terms of its estimated cost and the speed with which it could be accomplished.

The seven-month study prepared by a group of specialists headed by Graham, a former chief of the Pentagon's Defense Intelligence Agency, calls for a 10-year, $50 billion effort.

Graham argues that because the new project could carry out a number of communications and intelligence-gathering roles now handled by other systems, the money could come from "redirecting" funds rather than adding to planned budgets. He stressed that he was not suggesting the United States does not need new MX missiles and B1 bombers for retaliatory striking power.

The idea is to provide a "layered defense" against missile attack.

The first step would be a "cheap and simple" antimissile defensive ring around U.S. missile silos. It would send up a swarm of nonnuclear projectiles to blow up incoming enemy warheads about a mile away from their targets.

The study estimates this could be done in two or three years for about $1 billion and at least reduce Soviet confidence in a first strike. Pentagon officials believe this kind of missile defense is credible but point out that other concepts are already going ahead.

The next step, costing $10 billion to $15 billion and taking five or six years, would be a network of 432 satellites orbiting the earth, enough so that at least 25 percent would always be in position to intercept Soviet missiles soon after they are launched.

The satellites would carry heat sensors to detect the Soviet missile engine exhaust and would quickly fire 40 to 50 heat-seeking rockets carrying conventional rather than atomic warheads at the Soviet missiles. Such rockets are already being developed for another program.

While this concept has attracted attention before, Pentagon scientists say it isn't easy to do.

A final, $5 billion step 10 or 12 years from now and completing the "layered" system would place more technically advanced satellites and weapons in orbit that would be able to shoot down missiles midway in their flight path. The Pentagon is most skeptical of this.

Graham argues that this layered defense could reduce a Soviet attack "by 95 percent or better."