John A. Todhunter, who has said repeatedly that he did nothing to help his former employer while serving at the Environmental Protection Agency, last year tried to influence the Food and Drug Administration's plans to cancel a contract with the firm, according to government documents.

Internal memos show that FDA officials sought to cancel the contract because they regarded Todhunter's own work on the project as inadequate and below standard. But Todhunter asked an FDA bureau director to end the contract on terms more favorable to his former employer, Andrulis Research Corp. of Bethesda.

Soon afterward, Todhunter's office at the EPA began negotiations to give Andrulis a new contract for $40,000 without competitive bidding.

Rep. Mike Synar (D-Okla.), chairman of a House Government Operations subcommittee, has been investigating Todhunter's role in the Andrulis contracts.

"This new information raises serious new questions about Dr. Todhunter's continuing relationship with a former employer while he was in office at EPA," Synar said.

Todhunter said he had raised the Andrulis contract with Dr. Sanford A. Miller, director of the FDA's Bureau of Foods, "only in a general sense. I never asked Dr. Miller to do anything for Andrulis. I may have made some offhand comments . . . . I raised it as a background thing." Todhunter said the company's work was adequate, but that there had been some "miscommunication" with the FDA.

Todhunter resigned last Friday as assistant EPA administrator for pesticides and toxic substances.

Todhunter has said that while he suggested the need for the EPA contract he disqualified himself from any contact with the $40,000 award his office made to Andrulis. Todhunter sought and received a general counsel's opinion last April that the contract would be proper as long as Todhunter was not involved.

During the same month, the FDA was about to cancel a $54,686 contract with Andrulis. Todhunter had been the company's project director for the study of liver tissues. Because the FDA was dissatisfied with Andrulis' performance, the company stopped work on the contract in June, 1981, a month before Todhunter joined the EPA.

FDA officials said in internal memos that they were canceling the contract because "the contractor failed to submit acceptable data . . . failed to make the necessary corrections . . . failed to make progress" and had submitted "inaccurate or misleading" progress reports. One official, Debra Dugas, said in an interview that the firm's scientific work did not meet the contract's standards.

Her April memo also noted that Todhunter had called Miller and asked whether the agency could cancel the contract in a way that would be more favorable to Andrulis. Miller's staff passed on the request to the contracts office.

Dugas said the FDA wanted to issue a "termination for default," a negative finding that puts the blame on the contractor and may involve penalties. But Todhunter asked Miller for a "termination for convenience," which places no blame on the contractor and allows the firm to charge the government for some settlement costs.

The issue has remained in dispute since the FDA offered to pay Andrulis $13,000 for work already done, leaving the firm with an unpaid balance of $41,000. Andrulis, which is headed by a woman, received both the EPA and FDA contracts as a "set-aside" reserved for minority firms.

Todhunter recently amended his financial disclosure statement to show that Andrulis had paid him $9,421 for work as a consultant before he joined the EPA. Todhunter has acknowledged receiving $1,600 of that money after he joined the agency, saying this was deferred income for previous work.