The participation of Web2.0 in collaboration makes the Q&A social platform develop rapidly, but the problem that the answering quality is difficult to distinguish has gradually become apparent, even it hindered the healthy development of social platforms. In order to predict the Answers in Zhihu, which is the famous Community of Questions and Answers website, the quality prediction method based on Logistic Regression is proposed. Firstly, we collect dataset from Zhihu. Based on the factors, the training set is constructed. After that the two answer quality prediction models based on Logistic Regression is constructed. With the comparison of the two models, the final model is determined. It can be used to predict the answer quality of Zhihu.