A new report casts fresh doubt on the practicality of the Maverick, and the senator who asked for the study has urged that the controversial anti-tank missile be killed.

The General Accounting Office report says there is cause for concern about the safety of pilots who launch the heat-seeking "smart" weapon. The report also raises questions about how useful the missile might be in combat and questions a $442.8 million saving forecast by the Pentagon. It says evidence of this is "lacking or unconvincing."

The report was requested by Sen. David H. Pryor (D-Ark.) after the GAO said last June that five years of operational testing had failed to show that the missile "can be used effectively by U.S. military personnel in combat." The agency urged Defense Secretary Caspar W. Weinberger to hold back the program, but instead last September he approved "low rate pilot production" of 200 Mavericks.

Pryor has now asked the chairmen of the House and Senate Appropriations subcommittees on defense to consider halting funding. "The Air Force still cannot assure us that the Maverick will meet the test of the battlefield," Pryor said. "To take further production steps could be a costly and dangerous mistake."

The Air Force had no comment yesterday.

GAO said in its new report that "it is not possible to determine if . . . aircraft would survive" attempts to deliver the Mavericks over enemy territory at low altitude, at night, and in poor weather. This is partly because, to launch a Maverick, a pilot must fly a predictable straight path for perhaps seven to 15 seconds, more than enough time, critics say, for anti-aircraft gunners to shoot him down.

The GAO said the Air Force has neither addressed the issue of Maverick performance "in terms of the survivability of the aircraft" nor analyzed its data on this question.

In seeking funding in the mid-1970s for the heat-seeking Maverick, the Air Force claimed to Congress that it would be "a complete day/night/adverse weather system" and would complement an existing large arsenal of television-guided Mavericks intended for use only in clear daylight. Later, however, the Air Force gradually backed off day-night and all-weather claim and the GAO also questions it.

On adverse weather capability, the GAO said that developmental testing, "although planned for, was never successfully conducted." Testing in reduced visibility in simulated combat conditions has been limited, and "at this time there are no plans for an adverse weather assessment," the report added. The GAO also said that the program cost grew in less than a year from $4.9 billion to $5.8 billion and may grow even more if the predicted $442.8 million cost saving does not materialize from getting a second supplier to compete with Hughes Aircraft Co. and from a multi-year procurement plan.

The original 1975 acquisition goal, 31,113 missiles, was nearly doubled in 1980 to 60,697. The unit cost in the five-year period rose from $51,200 to $96,300.

Other highlights of the GAO report:

* The operational tests found Maverick software to be "deficient" as to "reliability, qualitative maintainability and . . . supportability." Of the 48 reports submitted by maintenance personnel in the operational tests, 22 were "for deficiencies that prevented the success of a mission . . . . By February, 1983, 10 of the 22 deficiencies had not been resolved. Twelve deficiencies were found to be causing marginal or degraded performance. Four remain open."

* In a sharp departure from war conditions, pilots in the operational tests became familiar with their targets. They "trained in the same area that the test missions were flown" and made as many as 15 "passes" at a target before firing a Maverick.