The report, by the General Accounting Office, also criticized the administration for refraining from making long-term cost estimates for many elements of the planned system, clouding decisions about what technologies to pursue.

The report echoed concerns that missile defense opponents in Congress and elsewhere have raised about the Bush plan. As the president has pressed toward his goal of putting missile interceptors in Alaska and California by September 2004, the political debate over missile defense has shifted from ideological arguments about arms control to practical considerations about performance and cost.

Critics accuse Bush of shortcutting normal Pentagon testing and budgeting procedures to have an antimissile weapon in place before the next presidential election. Administration officials contend that U.S. vulnerability to a ballistic missile attack warrants quick fielding of a less-than-perfect defense. They also say that tests so far have confirmed the viability of the basic concept behind the system: namely, launching missile interceptors into space to collide with enemy warheads.

But the GAO report said Bush's directive to build the system by 2004 has placed the Pentagon's Missile Defense Agency "in danger of getting off track early and introducing more risk into the missile defense effort over the long term."

Because of time pressures, the report said, the planned system contains components "that have not been demonstrated as mature and ready" for incorporation with other elements. One example cited was the three-stage booster rocket intended to lift the "kill vehicle" interceptor into space. Development problems have prompted the Pentagon to change prime contractors and order two booster designs, neither of which has been fully flight-tested yet.

In fact, the report noted, flight-testing of the whole system has remained crimped. Although interceptors have scored hits in five out of eight attempts since 1999, the tests have relied heavily on surrogate or prototype components and have been run under "non-stressing conditions" far different from those likely to be encountered in a missile attack, the report said.

"As a result, testing to date has provided only limited data for determining whether the system will work as intended in 2004," the report said.

The Missile Defense Agency issued a statement expressing confidence that the planned system would provide "an effective, reliable defense" of all 50 states. "This confidence comes from the outstanding technical success we have achieved in our development and test program," the statement said.

But two Democratic senators who have led congressional opposition to the administration's plan and who requested the GAO report -- Carl M. Levin of Michigan and Jack Reed of Rhode Island -- seized on the findings. "The GAO report provides a troubling picture of a system without direction," Reed said. "The president's decision to deploy an untested national missile defense system still seems to be motivated more by politics than effective military strategy."

Recognizing they do not have the votes in Congress to block Bush's initiative, Levin, Reed and other critics have concentrated instead on writing language into defense authorization measures that would force the administration to spell out performance criteria and operational test plans for the proposed system, and provide periodic assessments. The administration has so far resisted getting too specific about system requirements. While acknowledging that the system as initially deployed will be far from perfect, Pentagon officials say plans call for it to be improved over time in "blocks" as the technology matures and the threat evolves.

The GAO report said such flexible, phased development makes sense for a new, complex weapons system such as missile defense. But it faulted the Missile Defense Agency for not estimating the total "life cycle" cost of elements or budgeting for them. Agency officials agreed with this criticism and plan to start drawing up long-term cost estimates, the report said.