With the development of microblog, many studies pay special attention to sentiment classification of the reviews in microblog. This paper summarizes three well-known methods for text classification and then improves one of them for sentiment analysis. We come up with a new model in which we introduce efficient approaches to select features, calculate weights, train samples and evaluate classifier. The new model is based on Bayesian algorithm and machine learning that is one of the most popular methods for sentiment classification. Our model can enhance the overall efficiency of the sentiment classifier.