Federal health officials said yesterday that they may abandon plans for an extensive national survey to determine how many Americans have been infected with the AIDS virus because of widespread resistance to being tested.

Plans for testing 50,000 people across the country were announced last June after President Reagan asked the Centers for Disease Control to study the prevalence of acquired immune deficiency syndrome in the United States. But designing an accurate and comprehensive test has proved difficult, at least in part because up to one-third of those contacted have said they would not participate, even anonymously.

"The random national study is still under consideration," said Dr. James O. Mason, CDC director. "But I think the probability that it will be done on a national level is low."

State and local health officials met at CDC's Atlanta headquarters last week to discuss options for a survey. Many public health experts have said that without a broad national study there would be no clear way to determine how many Americans have been infected with the HIV virus, which causes AIDS.

The Public Health Service currently estimates that 1.5 million Americans have been infected, but that number is based on rough estimates made almost two years ago.

Throughout the summer, federal health officials in Atlanta and Washington debated ways to conduct such a survey. Although epidemiologists decided it was possible to conduct a national household sample, they concluded that it would fail to reach many of the people, such as drug users and homosexual men, at highest risk.

In addition, it soon became clear to researchers that assurances of confidentiality were not enough to persuade many of those at high risk for AIDS to participate.

"If somebody came knocking on your door and said, 'Hello, I'm from the federal government and I want to test your blood for AIDS,' what would you do?" asked Dr. Harry Hull, president of the Council of State and Territorial Epidemiologists. "This is not an easy problem to solve, but we really need those answers."

The CDC has begun several smaller surveys of key hospitals, sexually-transmitted-disease clinics and drug-abuse treatment centers. Thirty cities are participating in the current tests, 20 of which the CDC has judged as "high risk" areas and the others as "low risk" areas.

Early this summer, consultants told the CDC that a national survey of the type initially proposed would take up to three years. Many public health officials have said that was too long, and that a valuable small survey could be carried out much more quickly.

Legislators have called for wider testing to determine the prevalence of AIDS. Health officials have opposed many of the suggestions on the grounds that widespread testing would be unlikely to reach those at highest risk.

"What worries me most is that while we are arguing the whys and wherefores of AIDS testing in this country, the politicians are basically doing our jobs for us," said Dr. Alfred Saah, associate professor of epidemiology at the Johns Hopkins School of Public Health. "Screening 19-year-old brides in a suburb of Chicago is not the way to stop the epidemic or find out who is sick."

Saah said that an extensive -- and expensive -- study will be necessary to pinpoint with certainty the natural history of the disease. Most of those afflicted with AIDS so far in the United States have been homosexual and bisexual men, intravenous drug users and their sexual partners. A small but slowly growing percentage of those infected have been heterosexuals.