Iris is one of the most reliable biometrics because of its uniqueness and stability, hence has played an important role in many biometric applications such as access authentication. While existing Gabor features have demonstrated great success for Iris recognition methods, they are not designed for visually characterize Iris patterns. In this paper, we purposely design 6 Gabor filters after analyzing the texture and spatial information of the iris to extract features. However, the features may have redundant information and some non-effective features, which interfere with the matching process. To fuse the Gabor features obtained through these Gabor filters, we propose a weighted multi-view feature fusion algorithm by minimizing intra-class distance and maximizing inter-class distance. The Particle Swarm Optimization (PSO) algorithm is utilized to solve this nonlinear optimization problem. Experimental results on the popular benchmark dataset CASIA-IrisV4-Thousdand demonstrate that our proposed method utilizing the novel Gabor features and multi-view fusion algorithm outperforms other Gabor feature based methods.