An initial "Star Wars" deployment in the late 1990s might require about 600 space launches during a three-year period, more than one rocket liftoff every two days, an Air Force official said recently.

The official, Col. William Zersen, said that a first-line antimissile system might include 3,200 rocket-carrying battle stations in 40 different orbits and additional satellites carrying radars and other sensors. Zersen, assistant for advanced launch systems at the Air Force Space Division, disclosed the estimates at a San Diego conference that was reported in this week's edition of Military Space newsletter, although he stressed in an interview yesterday that he was discussing only preliminary "ballpark estimates."

After the failures of the Challenger space shuttle and the Titan and Delta rockets, the United States now has almost no space launch capacity. Yesterday, the Strategic Defense Initiative Organization, the Defense Department office in charge of the Star Wars research program, released an annual report that acknowledges the challenge facing SDI in space transportation.

Although the unclassified report contains few specific numbers, it notes that "success in nearly every element of the program is dependent on major advances in . . . space transportation and logistics," among other technologies.

"It is clear that there is not now an adequate knowledge of the supply requirements and logistics infrastructure to support a space force of the magnitude and complexity envisioned for a multitiered ballistic missile defense," the report says.

The annual report shies away from the previously stated criterion that any defensive system must cost less than the Soviet offensive system that might overwhelm it. The criterion, first enunciated by President Reagan's senior arms control adviser, Paul H. Nitze, held that any defensive system must be "cost-effective at the margin."

Lt. Gen. James A. Abrahamson, SDI director, said at a news conference yesterday that he is not trying to "sneak out" from that standard.

"It's the president's criteria, it's a logical criteria," he said.

But Abrahamson said that "it's more than merely an economic problem." He said many factors must be considered as the United States attempts to discourage the Soviet Union from increasing the size of its nuclear arsenal.

"Obviously, economics is an important part of that," he said. "But the Soviet Union is not a market-oriented economy."

The administration is fighting to increase SDI funding from this year's $3 billion to about $5.4 billion next year. Abrahamson said that if Congress substantially trims the request, as have key congressional committees, it will "substantially" and "dramatically" delay the program.

The report calls for major experiments and technology demonstrations in space and on Earth to provide a "timely, visible and understandable set of milestones."

But, in a response to a question, Abrahamson said he has not structured the program to respond to congressional pressure for tangible results. "That's not the way we're conducting the program," he said.