The Congressional Budget Office said yesterday that the little missiles President Reagan wants to build to supplement the MX blockbusters could cost more than $100 billion over 20 years and require 50,000 people to operate, repair and guard them.

The president's plan now being reviewed by Congress calls for building small, single-warhead missiles that would be hauled around six or seven Air Force bases in the West and, in a crisis, "flushed" onto public roadways.

These "Midgetmen" under Reagan's plan would be deployed around 1991, or three to four years after the MX. The 100 MXs proposed, each carrying 10 nuclear warheads, would go into existing Minuteman silos.

The president's advisory commission on the MX, in recommending development of the small missile, did not say how many would be needed, partly because the total would depend on how many warheads the Soviets had in the field.

The Air Force, however, has been making cost estimates based on a force of 1,000 mobile, single warhead missiles. The CBO used that assumption in its report.

Alice M. Rivlin, the CBO's director, said the "major message" in the forthcoming report on Reagan's plan is that the MX would add little to the U.S. nuclear deterrent if the Soviet Union struck first. Although the CBO in its summary report issued yesterday doesn't judge Reagan's strategic program, it is heavy on findings that could be exploited by critics of both the small missile and the MX.

The findings included these, with costs expressed in fiscal 1984 dollars:

* Small, mobile missiles. "Fielding 1,000 such missiles would cost $46.2 billion to buy, plus about $3 billion in annual operating costs, for a 20-year life-cycle cost of $107 billion." Larry Cavaiola and Bonnie Dombey, authors of the report, said that those were Air Force estimates expressed in fiscal 1984 dollars. CBO officials also said the 50,000 people needed to operate, maintain and guard the 1,000 missiles around-the-clock is also an Air Force estimate.

A separate Air Force assessment distributed by the CBO yesterday states that "a continuously road-mobile system has very high survivability. Its military capability is not as good as most other alternatives since it has limited hard-target capability," which could be improved if deployment were delayed two to three years. "This road-mobile system relinquishes the protection offered by a secure area on a military reservation for major exposure to the public. It is manpower intensive . . . . "

* 100 MX missiles. "A major Soviet strike that the United States rides out could destroy all but 10 percent of them in 1990 and all but 5 percent by 1996 as accuracy continues to improve. This limited survivability, coupled with the major buildup in other U.S. strategic forces, means that deploying the MX would contribute little to U.S. retaliatory capabilities.

"Specifically, 100 MX in Minuteman silos would contribute 3 percent in 1990 and less than 1 percent in 1996 of all U.S. hard-target warheads that would be expected to survive and be available for retaliation in the nuclear scenario considered most likely--a Soviet first strike occurring after some warning."

* Canceling the MX would save $18.4 billion. Alternatives to building that missile include turning more Minuteman 3 missiles into silo busters for a cost of $14.4 billion and deploying five to nine more Trident submarines to substitute for nuclear firepower of the MXs that would survive an attack.