Stock trend prediction problem is an attractive research topic since many factors should be considered simultaneously. Due to the fact that financial news articles are a combination of influential information and noises, extracting valuable insights in between noises is an interesting and rewarding task. Hence, lots of approaches have been proposed for stock trend prediction based on financial news. However, they focused on analysis of English news. In this paper, we thus first depict the Chinese news-based stock trend prediction framework. Based on the framework, we then propose the stock trend prediction (STP) algorithm that takes Chinese news and technical indicator into consideration for predicting stock trends. Because parameters setting is an optimization problem, we then modify the framework and design the second stock trend prediction algorithm to determine the optimal trading situation using genetic algorithms, namely GATSP. At last, experiments were conducted on real news and stock data to verify the effectiveness of the proposed approach.