In the field of face recognition, the application of sparse representation is very successful, even if the images are taked under different illumination or including facial expression, it still has good recognition effect. However, when the test and training images contain both the changes of illumination and expression, the traditional sparse representation algorithm often performs a wrong face recognition. In sparse representation, the l1-norm was used to define the fidelity of sparse coding. And the face feature extraction based on sparse representation is too simple, and the sparse coefficient is not sparse. In this paper, we propose a simple and effective face recognition algorithm, in which the classification algorithm sparse representation and Gabor feature are fused effectively. The useful information of feature can be fully reflected in the sparse representation. Hence the new residual values, that are obtained, can improve the fidelity of residuals. We exploit the fusion nature of sparse coefficients to redefine the computing method of residuals, and then perform classificcation. We conduct several experiments on publicly available database to verify the efficacy of the proposed approach and corroborate our claims.