The Air Force, already struggling with delays in the deployment of its highly touted new MX strategic nuclear missiles, has decided to suspend all MX flight tests and empanel a special study group to resolve new questions about the missiles' accuracy, according to Defense Department and congressional sources.

The flight test suspension came to light as the House Armed Services Committee blasted the Air Force in a new report for poorly monitoring the MX program and failing to notify Congress of serious problems in the guidance system of the missile that Air Force officials have called "the keystone of our nation's strategic defense."

"As a result of. . . significant management deficiencies and ineffective program monitoring, serious questions of confidence" have arisen in the 21 MX missiles already deployed in Wyoming silos, the committee said in a report released yesterday after a five-month investigation.

Brig. Gen. Charles A. May Jr., the Air Force deputy director for advanced programs, called this assessment misleading and defended the Air Force's management of the MX program. "We really don't think there's a horror story here at all," he said in an interview.

The committee had previously disclosed that only 14 of the deployed MX missiles are considered "on alert," or ready for wartime use, while the remaining seven currently lack guidance systems needed to ensure they can hit Soviet targets with pinpoint accuracy.

The reason is that the Northrop Corp., which makes a key component of the guidance system, fell behind schedule in its deliveries to the Air Force, which suspended some payments to the company last year and initiated a criminal investigation of the company's activities.

But the committee called into question the reliability of even the 14 "alert" missiles by noting that Air Force estimates of their probable accuracy were drawn mostly from flight tests that used different versions of the guidance system. Moreover, the "trend in the later test flights has been toward less accuracy," the committee said.

"We're pretty damn happy with how accurate {the missile} is," May responded. But he confirmed that the Air Force decided to delay the three remaining MX flight tests "so that we could better understand these changes" in missile accuracy as the tests became increasingly realistic. It also asked a panel of its Scientific Advisory Board to initiate an "in-depth technical review" of the guidance system.

Although Air Force officials have long based their predictions of the missiles' unerring accuracy on data from the 17 flight tests conducted to date, the Air Force Operational Test and Evaluation Office told the House committee that only three of the tests are "operationally representative flights" because others did not use computer programs or equipment identical to that installed in deployed missiles.

The accuracy of the MX missile, which carries 10 warheads, is widely considered its most important military characteristic. Pentagon officials have stressed that Soviet efforts to encase their missile silos in highly protective concrete and bury their command posts underground demand the deployment of a U.S. offensive missile with unerring accuracy.

"At this point a giant question mark hangs over the MX," said Rep. Les Aspin (D-Wis.), the House Armed Services Committee chairman.

Rep. Samuel S. Stratton (D-N.Y.), chairman of the military procurement subcommittee, added that "We've got a case of deja vu."

"A few months ago, we found major problems with the B1 bomber that the Air Force had papered over for years," Stratton said. "Now we find problems with the MX that go back years."

Committee members were said to be particularly incensed that the Air Force had assured Congress in a December 1985 report that a "schedule variance," or delay, in Northrop's production of the guidance components would cause "no program or contract impact."

Only upon investigation did the committee learn that three months earlier, Maj. Gen. Aloysius Casey, a commander of the Air Force Ballistic Missile Office, had written to a senior Northrop official that the company's "inability to establish and then live up to reasonable delivery dates has caused grave mission impacts and, therefore, endangers . . . deployment" of the MX missile.

Casey's letter also called the delays a "chronic problem," and said they raised questions about the company's overall ability "to manage and perform under the contract." He explicitly threatened to assess penalties and withhold future contracts from the company if the problem was not corrected.

After reviewing additional Air Force documents, which brushed off some of these difficulties and discouraged a wider Pentagon investigation, the committee concluded that the Air Force put "a premium . . . on pushing components and missiles out the door as quickly as possible and on minimizing any program difficulties."

But May responded that "there was no attempt on the part of anybody in the Air Force not to tell people {about the problems} at what we thought was the proper level."

Both the Air Force and the committee agreed that many of Northrop's troubles stemmed from the swift transition from research to production, which necessitated hiring 4,500 new employes in a brief period.

To meet a tight schedule mandated by both Congress and the administration, Northrop secretly established a network of "dummy" companies to buy parts for the guidance system outside of approved channels. Similar fictitious companies also have been used by Northrop without approval to buy parts for the Air Force's multibillion-dollar Stealth bomber program, the committee said.

Parts acquired through these companies were "not necessarily inspected and certified," the committee said, and some important test results were falsely certified.