The National Academy of Sciences yesterday recommended that the U.S. Postal Service move quickly to replace about one-half of all first-class mail with an electronic system that would eventually put message terminals in homes and offices.

In a 51-page study that was sometimes cautious and sometimes highly critical of the postal service, the academy urged "a firm and continuing commitment to involvement in the electronic message field."

The report, presented to a commission studying postal problems and finacing, said that "electronic message systems are neither a panacea nor a guaranteed solution to the present problems of the postal service," but that they might be able to significantly reduce or eliminate deficits such as the $1.2 billion the postal service lost last year.

A spokesman for the postal service, which funded the report, said that the service was studying it and for the moment had no further comment.

The Commission on Postal Service, established late last year and told to make recommendations by March 15, will conduct the last of five hearings here today amidst growing signs that Congress may seek to restore more direct controls over the nation's post offices.

"Time is running out for the United States Postal Service," the report continued, "While it delays its entry into electronic message services, developments by private firms are sure to proceed, possibly foreclosing any opportunities for the USPS to move into the field in any meaningful way."

The report criticized the postal service for devoting "less than one-half of once per cent of its operating budget to research and development," compared with an average of 5.2 per cent for the electrical and electronics industries.

The report recommends a three phase approach that in its final stage would allow a sender to communicate directly with a receiver through electronic terminals in their homes or offices.

According to the study, whose recommendations are meant to be implemented between 1985 and 1990, all of the technology is either currently available or in some stage of development.

With respect to precise cost, it says, "There are too many factors involved to determine the economic feasibility of the service." It adds that the post office may not benefit economically from the electronic service until the third phase is operating, but that the first two phases must be put in place, at least in prototype, before the third can be set up.

In the first phase mail would be handled conventionally until it reaches a bulk handling facility. At that point the message would be transmitted electronically to the bulk handling facility nearest the addressee. There it would be converted to print form and delivered conventionally to a local post office and finally to the designated home or office.

The second phase would work much as mailgram service does now, with the sender initiating the message in any oof a number of ways, including over the telephone, and a printed copy being delivered to the home or office of the addressee.

Phase three would be pure electronics, with the sender initiating the message on a terminal in his home or office and the addressee receiving it on a similar piece of equipment, which may or may not produce a printout.

Systems of this kind are is use already, especially in the banking industry. New York's Citibank has investee $100 million in electronic systems, according to the report.