With the advent of the digital age, information has become more and more important, and people’s identity information has also received more and more attention. Iris recognition is an efficient and convenient identification technology, which has been applied to mobile phone payment, public security, personnel attendance and other scenarios. However, as iris recognition has become more popular, a large number of small-scale iris recognition have emerged. Faced with this situation, the training samples provided to the neural network class method are not enough, and the accuracy of the simple range-class method is too low. In this regard, this paper proposes an iris recognition method based on asynchronous genetic particle swarm algorithm (GAHPSO). The algorithm uses the embedded hybrid algorithm framework for reference, and selects appropriate crossover operator, mutation operator and possible occurrence probabilities according to small-scale iris recognition problems. In addition, it selects inverted S-type inertia weight and asynchronous learning operators to replace constant values. The optimization ability of the algorithm is improved. On the basis of distance class method, Gabor filter is optimized by asynchronous genetic particle swarm algorithm, which enhances its ability to extract iris features without changing the structure. The experimental results show that the proposed method can accurately and quickly carry out small-scale iris recognition, which is better than other existing schemes.