Some of the nation's leading specialists in human behavior have told the Secret Service it needs to improve its ability to identify potential assassins in what many call "a time of increased risk" for a U.S. president.

Among other important steps, said the behavioral scientists, the service could improve its interviews with possibly dangerous persons to yield more telling data, train its agents to understand mental aberrations and modernize its information system--all to help identify possible assassins and high-risk behavior patterns.

This advice is about to be published by the National Academy of Sciences in a summary of an unpublicized conference last March between high Secret Service officials and psychiatric and behavioral experts assembled by the academy's Institute of Medicine.

The conference was held just three weeks before the attempt on the life of President Reagan.

In an opening talk, Robert A. Snow of the Secret Service Office of Protective Research called it an "unprecedented" and "historical" cooperative effort between behavioral scientists and the president's protectors, organized at the service's request.

Secret Service spokesman Jack Warner said last week that "we've just received the report and need to study it," but "I'm sure we're going to pursue" these problems "and try some of these suggestions."

In three days of meetings, the scientists largely agreed that the Secret Service has done a good job but needs to do still better at what Dr. W. Walter Menninger, conference chairman, called "the impossible job" of fully protecting public officials.

Menninger, a senior psychiatrist at the renowned Menninger Clinic in Topeka, Kan., said in an interview last week that there are so many growing elements of risk in our very way of life today that "we can't count totally on the Secret Service" to eliminate them.

Included, he said, are "our high mobility, our easy access to guns, our focus on individual rights" and today's "emotional, single-issue, highly polarized political climate which tends to emphasize divisions in people instead of bringing people together." Also: "One of the potential costs of the cutback in government spending is that we're going to have a number of aggrieved persons who feel they've been injured."

These are things society, not the Secret Service, must deal with, he said, but they also mean "the Secret Service needs help."

The service keeps regular track of between 250 and 400 potentially dangerous persons, interviews them every three months and constantly interviews others in its growing file of some 26,000 persons who at some time have threatened or acted threateningly toward political leaders.

Yet, Menninger said, "there is no consistent interview format," no complete list of key questions "so one is sure you've got all the data you need," and "there are some times when information is missed."

Also, he said, the service's promotional structure means "people don't specialize." So a good interviewer and judge of possible "dangerousness" might not stay in that job, and an agent who gets the job might not be good at it.

The conference produced no formal "recommendations." But the consensus, said Menninger, was that the Secret Service should:

* Develop an improved interview "format" to help identify the truly dangerous, and also lay a data base for regular use and for research. "Not only is there a lack of standardized reporting for such things as psychiatric and criminal history of threateners," Menninger says in the official summary, but the service's research capacity, "its ability to learn from its experience and determine whether its agents are operating" efficiently, is "extremely limited."

* Give agents training in dealing with, and even helping, the mentally ill, and distinguishing between dangerous and harmless conditions, since 90 percent of the persons who threaten politicians have histories of mental illness. "In effect," said Menninger last week, "the service came across as a mental health service, because they're so often dealing with people who need mental health help. It's obvious that some of the contacts the service gets are distorted cries for help."

* Develop relationships with mental health agencies and practitioners from whom the service can get consultation and assistance, and also establish a "behavioral science research advisory group" to help develop better ways to identify dangerousness.

* Develop a computerized "decision support system," a computer program that compares all the characteristics of a suspected individual with known "risk factors." Often such a model can prove more reliable than any expert, Menninger said.