The Army said yesterday it may have to plow through 12 million microfilmed documents seeking the names of soldiers exposed to radiation during nuclear tests in the 1950s because it lost the identification forms the soliders filled out in quintuplicate at the time.

While Army officials were detailing the painstaking efforts they are going through to locate soldiers involved in the tests, Rep. Paul G. Rogers (D-Fla.) produced a letter from the Defense Department written in 1970 which stated that at least 80 percent of the soldiers had already been identified and their records placed on permanent file.

Rogers heads the House subcommittee on health and the environment, which is looking into the aftereffects of nuclear radiation on servicemen and others involved in the tests.

An estimated 300,000 military and civilian observers were exposed to radiation during 307 nuclear explosions set off in Nevada and the Pacific Ocean between 1945 and 1962.

The investigation into the tests was touched off by reports that several of the servicemen involved in maneuvers around the nuclear explosions have since developed Luekemia. One, former paratrooper Paul R. Cooper, who took part in the 1957 "Operation Smoky" maneuver in Nevada, died of Leukemia in Boise, Idaho, last week.

Military spokesmen said yesterday that it may take two or three more years to determine the full extent of the cancer risk to those who were part of the tests.

The Army is apparently having the most difficult problems in its search for test participants. At least 50,000 soldiers took part in 51 series of maneuvers which coincided with nuclear blasts at Camp Desert Rock in Nevada.

Army spokesman Col. Victor J Hugo said that security records filled in by each soldier at the time of the tests have been lost.In their place, Hugo said, Army officials are going through 12 million microfilmed unit morning reports to learn the names of individuals assigned to the tests.

Rogers read a letter written in 1970 from the Defense Department to Rep. Olin E. Teague (D-Tex.) stating that 80 percent of the troops involved in the tests had been identified and that none had received "a medically significant exposure."

Hugo said the records may be part of a file which was turned over to the Department of Energy but which has not been released because of the federal privacy act. An Energy Department official said the records would be made available in 30 days.

Rogers also criticized the Energy Department for falling to provide follow-up investigations on its former employes who received relatively low doses of radiation on the job. The Florida congressman noted that several other witnesses before the committee testified that low radiation doses may cause cancer.

One of those witnesses, biostatistician Irwin B. Ross of Roswell Park Memorial Institute in Buffalo, complained earlier yesterday to the committee that his research on low-level radiation and that of others has been suppressed recently by what he called "big science."