To enhance the accuracy of the text classification, a new method based on quantum PSO and RBF neural network is proposed. Firstly, it establishes the key words set to describe the classification of the samples, and uses fuzzy vector space model to build the feature vectors of every kind of sample, then automatically classifies the texts by RBF neural network, optimizes the parameters of RBF neural network by improved quantum PSO to enhance its approximation capability. The new method is proved by the classification of some documents in China periodical document database. The experiment shows that this method makes significant improvements in classification accuracy compared to other methods.