For the first time since the FBI began compiling crime statistics 25 years ago, crime has decreased for three consecutive years.

The number of violent crimes reported to police across the nation remained unchanged last year, but a decline in property crimes, such as auto theft and burglary, caused the number of total reported crimes to decline by 3 percent last year, according to figures released by the FBI yesterday.

Most of the decline came from rural areas and cities of less than 250,000 population.

In Washington, which is one of the 175 large cities that the federal agency tallies separately, the results were positive. Most violent crimes showed a decrease, while reports of auto theft increased 9.5 percent. Overall, reported crimes in the District dropped by nearly 8 percent, according to the FBI.

According to the FBI statistics, there were 175 murders reported to District police in 1984, a 4 percent decline from the previous year. There were 366 rapes last year, which is 40 fewer than were reported in 1983, according to the agency's figures.

Robbery showed the biggest decline among the violent crimes in the District, with 1,611 fewer incidents last year, or nearly a 21 percent drop.

The number of burglaries in the District also decreased. Police reported there were 10,954 burglaries in Washington last year, 1,529 fewer than in 1983.

A D.C. police spokesman said yesterday that the department will have no comment on the findings.

The nationwide drop in crime has been attributed to the aging of the post-World War II baby boom generation, as well as an increase in the number of people in jail.

In 1982, serious crime dropped 3.4 percent in the United States. In 1983, serious crime dropped 6.7 percent, the largest decline since the FBI began compiling crime figures.

The figures are based on crimes reported to nearly 16,000 police agencies across the country.

The FBI's crime survey has drawn criticism because it counts only crimes reported to the police. A study by the Police Foundation, a research group, about a year ago found that local arrest statistics often were inaccurate because local police did not follow the bureau's complex instructions.