Facial feature extraction is the most important step in face recognition process, and the quality of the feature directly affects the recognition rate. In order to get better face recognition effect, it is necessary to make full use of sample information. To make full use of the information contained in training samples and test samples, we propose a semisupervised linear discriminant analysis (SLDA) feature extraction method based on weighted combination of principal component analysis (PCA) and linear discriminant analysis (LDA). At the same time, inspired by the combinatorial optimization problem, we use the binary genetic algorithm to optimize the feature space obtained by the semisupervised feature extraction method. Experimental results of ORL face database show that the proposed method achieves higher recognition rate than the classical face recognition algorithm and some improved methods.