A panel of university and defense officials has agreed on proposed steps to protect scientific research from disclosure that might harm national security, panel co-chairman David A. Wilson told an audience of scientists today.

It is the first formal agreement on the subject between academics and the government and, although not binding, is expected to influence the final policy now being developed by the Senior Interagency Group on the Transfer of Strategic Technology, an independent body reporting to the White House.

The compromise was reported at the meeting of the American Association for the Advancement of Science here.

Officials from Stanford, Massachusetts Institute of Technology, the University of Minnesota, the University of California system and other universities agreed to two types of controls on "sensitive" research done under government contract, most of it for the Defense Department:

* A government contract officer would have the right to review a research paper for 60 days before publication and could try to persuade the scientist or his university to change it or not to publish it.

* No foreign citizens from communist or other named countries could participate in the research without government approval.

Universities had been reluctant to concede the second point, because universities would have to police some students and researchers to be certain that they had no access to labs doing "sensitive" work.

A debate between academics and government officials has been going on for several years since the government first attempted to crack down on publication of research that government officials feel might be militarily useful to the Soviet Union and other Warsaw Pact nations.

Three years ago, the National Security Agency slapped secrecy orders on some university researchers working on the mathematics of codes, only to rescind the orders later after the researchers protested.

Last year, the government physically seized more than 100 papers from a scientific meeting in San Diego because they covered such topics as laser guidance systems and, government officials said, Soviets and other Eastern bloc representatives were present.

Bobby R. Inman, then deputy director of the CIA, warned scientists last year that if voluntary controls were not put on the "hemorrhage" of technology and technological information leaking to the Soviets, even more restrictive legislation might be enacted.

However, the National Academy of Sciences issued a report saying that government and intelligence officials could not produce a single instance in which university research was leaked to the Soviets and used to military advantage.

The report acknowledged that the Soviets had gained access to much western technology and put it to use in weapons, but it said almost all the "technology transfer" came in the form of hardware sold legally and illegally to communist nations.

The report released today by the defense-university panel endorsed the conclusions of the National Academy of Sciences report.

According to panel chairman Wilson, some Pentagon officials were displeased with the new report, and the discussions at times became so heated that the university officials threatened to walk out of one meeting.

Eventually, he said, dissenting Pentagon officials set up their own group to recommend tougher controls, a study which still is going on.

Louis T. Montulli, senior policy analyst with the President's Office of Science and Technology Policy and one of the officials working on the Senior Interagency Group, said the proposed steps will be considered, but added that university research represents only a small portion of the problem of leaks to the Soviets.

He said leaks of some government research, however, is a serious problem. Montulli said that of the last 50 research papers logged into the NASA computerized bibliography, 35 were defense-related and might be senstive.

One, he said, concerned the performance of the F16 fighter's wing during close combat.

Wilson said that most university research would not come under the proposed controls, but that most government research would.