WASHINGTON - Nuclear experts said it is conceivable that only two people died in the Soviet nuclear accident, as Moscow claims, and U.S. intelligence information may indicate a worst-case scenario that is not occurring.

They said the official Soviet death tally of two is more plausible than unverified reports repeated by various government officials that as many as 2,000 may have died.

The experts, as well as members of a government task force, also called into question earlier suggestions from unnamed U.S. officials who a day earlier strongly suggested that two reactors had been involved in the accident.

Reagan administration sources, citing intelligence information, have postulated a disastrous chain of events involving a meltdown, chemical explosion and nuclear fire at one reactor and said there were indications of similar problems at the second reactor.

And virtually every official commenting on the crisis has dismissed the official Soviet death toll.

"I imagine the Russians want to put the best face on it. On the other hand, there are those who've said this is the worst thing that could have happened under the worst circumstances. The truth probably lies somewhere in between," said Frank Graham, vice president of the Atomic Industrial Forum.

"There are a lot of things that indicate it may not be the worst case," he added, citing relatively normal activity going on in Kiev, 80 miles south of the reactor complex.

Tom Cochran, senior staff scientist with the Natural Resources Defense Council, an environmental watchdog group, said he did not know whether the U.S. intelligence community was overstating the seriousness of the accident.

He added: "There are a number of factors that suggest that it could be very serious. But they could be consistent with something much less serious, given what limited data we have."

Graham and Cochran, while saying they suspect fatalities were higher than two, called the Soviet figure plausible and "within the realm of credibility." Alan Krass, senior arms analyst for the Union of Concerned Scientists, a nuclear safety watchdog group, also called the low number conceivable.

Krass and Cochran both said 2,000 casualties - an unconfirmed estimate widely quoted on Capitol Hill - was a more preposterous number than two.

"Radioactivity doesn't kill that fast. It's not an instantaneous killer like at Bhopal with the poison gas," said Krass.

Krass said the satellite pictures and infrared images on which U.S. intelligence analysts are basing their conclusions are not definitive. The unique situation makes it even harder to interpret the images, he said, because there is nothing to compare them with.

Presidential spokesman Larry Speakes and a U.S. task force monitoring the accident both said Thursday that they could not confirm reports of a meltdown or fire at the second reactor - although administration sources had strongly suggested Wednesday that was the case.

"The second hot spot seen in the LANDSAT photos is not a reactor," the task force said in a statement. Speakes said it could be another building burning.

Cochran said a Soviet photograph of the plant after the accident showed no smoke or flames coming from the second reactor building. "I have a problem," he said, contending either U.S. intelligence is wrong or the photo was doctored.

Graham said the fate of the second reactor is still highly speculative. "If you want to be optimistic, you can say it wasn't really so bad at (the number four reactor), they're forgiving machines, so probably number three is all right."