The invention discloses a face sketch generation method based on subspace ridge regression, and mainly solves the problem that an existing method is low in the quality of the generated sketch or is high in time consuming for sketch generation. The implementation steps of the face sketch generation method based on subspace ridge regression include the steps: 1) dividing a training sketch sample set, a training photo sample set and a test sample set; 2) dividing all the images into image blocks and forming a block set; 3) dividing the training photo block set and the corresponding sketch block set into a plurality of sub sets; 4) calculating a mapping coefficient matrix between every pair of photo-sketch block sub sets; 5) dividing the test photo block to the corresponding sub set; 6) according to each test photo block sub set and the coefficient matrix corresponding to the sub set, solving composite sketch block sub sets; 7) combining the composite sketch block sub sets to obtain a composite sketch block set; and 8) combining all the composite sketch blocks to generate a pseudo sketch. The face sketch generation method based on subspace ridge regression has the advantages of being high in the quality of the composite sketch and being high in the speed, and can be used for face retrieval and identification in the public security field.