A federal judge here has dismissed a suit by the Community for Creative Non-Violence over a study by the Reagan administration that showed there are 250,000 to 300,000 homeless people in the United States -- a fraction of the group's estimates.

U.S. District Court Judge Thomas Penfield Jackson held that the highly controversial report, issued in May by the Department of Housing and Urban Development, was not subject to rules governing some agency actions that can be challenged in federal court.

Mitch Snyder, director of the group which filed the lawsuit in June, said yesterday that Jackson's decision will be appealed. CCNV has estimated there are 2 to 3 million homeless people nationwide.

"Contrary to Judge Jackson, the report does impact on people and, come winter, it will kill people," Snyder said. "Judge Jackson is a Reagan appointee and we're not surprised that he upheld the government."

In its complaint, CCNV said the report "wrongly minimizes the number of homeless people and trivializes their plight."

CCNV, a Washington-based, nonprofit organization that provides services to the homeless, had asked the judge to force HUD Secretary Samuel R. Pierce Jr. to disavow the study's results and withdraw the report.

In dismissing the lawsuit, Jackson said CCNV had alleged "nothing to suggest that the level of public funding or private charity they expect to be forthcoming to meet the needs of the homeless" was likely to be affected.

"We're delighted," said June Koch, assistant secretary of HUD for policy development and research. "I can't say it was frivolous, but the suit and what's behind it tend to deflect the focus from the real problem . . . . I think Snyder's motives are high motives, but I disagree with his numbers."

Pierce has defended the report as the most scientific survey of the homeless ever conducted. Pierce derided estimates by CCNV as "guesswork."

The report, which drew sharp criticism from Capitol Hill liberals, put the number of homeless in the District at between 3,000 and 6,400. Snyder maintained at the time that 5,000 to 10,000 was more realistic.

HUD, conceding that the temporarily homeless are difficult to count because of a high turnover rate, said the study relied on interviews with 500 observers in 60 urban areas, national surveys and site visits.