The Carter administration asked Congress yesterday to restrict genetic research by licensing scientists who combine genes to make new forms of life.

Health, Education and Welfare Secretary Joseph C. Califano Jr. told the Senate Health Subcommittee that such a law must be enacted to protect the public from unanticipated epidemics or other dangers.

A genetic research licensing law, if enacted, would be the nation's first ban on basic scientific research that may be dangerous.

Califano said states and cities should be allowed to set even tighter gene research safety standards than federal rules although such a policy would be contrary to most federal-state relations. For example, on March 29 the Supreme Court ruled that the federal Wholesome Meat Act, the Food, Drug and Cosmetic Act, and the Fair Labeling and Packaging Act all pre-empt local standards in the interest of nationwide uniformity.

Califano thus put an administration stamp of approval on the kind of action taken by the Cambridge, Mass., City Council in February. In setting rules tougher than the federal guidelines it said neither Harvard nor Massachusetts Institute of Technology may perform the most potentially perilous gene-splicing experiments - the kind that will be done in "high containment" laboratories now being prepared by the National Institutes of Health on its Bethesda, Md., campus and at the former Army germ warfare center at Ft. Detrick, Md.

Most scientists, though agreeing with the need for federal safety rules for such research, have urged that there be one rule for all. Otherwise, they have said, "chaos" may develop, with scientists moving from place to place to find congenial rules - and some universities, and probably scientific progress suffering.

A federal interagency committee last month backed this view. So did the first version of an administration bill circulated in Congress a few weeks ago.

But Califano yesterday came around to the local-control point of view urged by the chairman of both congressional health subcommittees, Rep. Paul Rogers (D-Fla.) and Sen. Edward M. Kennedy (D-Mass.).

Most scientists and the interagency group have urged that both universities and private industry be governed by a single set of safety guidelines much like the NIH guidelines that now govern all recipients of federal grants.

Califano agreed, saying that the tougher rules, whether federal or state, should apply to both academic and industrial genetic research.

Genetic research may lead to new drugs, new farm products and new understanding of nature, but its dimly understood hazards leave "no reasonable alternative to regulation," Califano argued.

"While such regulation runs counter to my own commitment to the principle that intellectual inquiry can thrive only when free of constraint, I am persuaded," he said, "that with the use of recombinant DNA techniques" - techniques joining the deoxy-ribonucleic acid, or genetic material, of different organisms - "caution and restraint is required," at least until more safety testing determines whether such efforts are dangerous.

Many laws and rules now limit the kind of tools scientists may use. For example, licenses are required for use of dangerous radio active materials.

But with the proposed genetic regulation, warned Dr. Norton Zinder of Rockefeller University, "we are moving into precedent-making . . . regulation of an area of scientific research, and, I must plead that this be done with extreme care and without haste.