Less than a year after its first attempt went down in flames, the Reagan administration is preparing another swipe at the concept of "worst-case" analysis in gauging the environmental effects of major federal actions.

Under a draft proposal drawn up by a task force of lawyers from at least four Cabinet departments, agencies no longer would have to consider the impact of catastrophic accidents -- such as a loaded oil tanker breaking apart in Galveston Bay -- unless such events are "reasonably foreseeable."

The proposal would significantly limit the reach of the National Environmental Policy Act, under which federal agencies must prepare detailed analyses of the potential impacts of projects ranging from dam and highway construction to offshore oil leasing and gypsy moth controls.

The White House Council on Environmental Quality floated a similar proposal last year, arguing that agencies were wasting time and money trying to figure out the impact of events that "may well be highly remote or unlikely."

That proposal was offered as an advisory to agencies rather than a formal regulation, but it drew a flock of complaints from environmental lobbyists and their congressional allies. The CEQ quietly withdrew it six months later.

CEQ Chairman A. Alan Hill acknowledged this week that the council intends to reopen the case soon, by issuing a request for suggestions on how the administration might rewrite NEPA regulations to "more clearly define" the use of worst-case analysis.

But the council already has at least two proposals in hand -- from a task force headed by the Justice Department and from the Pacific Legal Foundation, a nonprofit law center that specializes in conservative causes.

Both proposals would go significantly farther than the advisory CEQ offered last year. The Justice proposal, for instance, would limit worst-case analysis to events that are "reasonably foreseeable," and only if the costs of obtaining the necessary information "are not excessive."

Moreover, the proposal says that an environmental impact statement "need only forecast those consequences that have a basis in credible scientific evidence, rather than consequences that are purely hypothetical or conjectural."

The Pacific Legal Foundation's proposal apparently would be even more restrictive, limiting worst-case analysis to cases where "a reasonable possibility of the occurrence . . . has been proved upon a firm evidentiary basis in the actual circumstances of the contemplated project" and a method for assessing the consequences "has been established."

Hill denied that the CEQ has any specific proposal in mind, even though the Justice document arrived at CEQ's offices early this month already typed up in the formal style of a Federal Register notice.

Hill called the document "a consensus proposal which we will consider with all of the others."

Roger J. Marzulla, acting deputy assistant attorney general for the lands division, who represented Justice on the task force, also played down the significance of the document. "The call has to be CEQ's," he said. "We just tried to make it easier for them."

But other administration sources said they expect the task force participants -- Interior, Agriculture, Energy and Justice -- to press for far-reaching changes in the NEPA regulations as a result of a recent Supreme Court decision not to hear a case involving the U.S. Forest Service's herbicide-spraying program.

The court let stand a decision favoring a group of Oregon activists who argued that the Forest Service had not made a worst-case analysis of the chemicals' potential health hazards.

Up to now, relatively few NEPA lawsuits have involved the concept of worst-case analysis, but the administration is concerned that the Oregon group's success will encourage similar challenges.

Forest and range management programs at Agriculture and Interior, for example, rely heavily on chemical pesticides and herbicides, many of which have not been fully tested for their long-range effects on human health.

The administration's new assault is certain to infuriate environmental and public-health groups, which have vigorously defended worst-case analyses as valuable exercises in planning for health and environmental protection. The Nuclear Regulatory Commission, for example, didn't require nuclear power plants to have evacuation plans until after the 1979 accident at Three Mile Island in Pennsylvania.

In a recent letter to Hill, two members of the Senate Environment and Public Works Committee called attention to a more recent example -- the Dec. 3 gas leak that killed more than 2,000 neighbors of a pesticide plant in Bhopal, India.

That tragedy "underscores the fact that improbable but catastrophic events do occur," Sens. David F. Durenberger (R-Minn.) and Jennings Randolph (D-W.Va.) wrote.

Hill said the CEQ is "thinking about that . . . . That's a case where you have a situation with low probability but massive impact. That's a classic case."