Abstract:Natural language processing technology has a very broad application prospect in people’s lives. Statisticalbased machine learning methods have become the mainstream technology in present natural language processing, in which various support vector machine techniques have been widely used. In order to improve the accuracy of text and sentiment classification in natural language processing, a machine learning method based on particle swarm optimization support vector machine is proposed. The method optimized the parameters of the support vector machine by constantly updating the position, speed, and optimal position of the current particle, so as to find the best support vector machine. Experimental results of text and sentiment classification show that the proposed particle swarm optimization support vector machine method has good performance in terms of accuracy.