[Objective] To explore the feasibility of the K-means clustering method in the classification of farmland types in Heilongjiang Province and provide support for the planning and management of soil and water conservation in slope farmland.[Methods] We made a clustering of farmlands in 70 cities(or counties) in low mountain and hilly areas in Heilongjiang Province using K-means clustering method.[Results] The cities or counties in the low mountain is 21 and hill area is 49.[Conclusion] The 16 classification indexes were significant at the level of 0.05, indicating the fitness of classification.