Recently China has been in a high-incidence season of food safety events and not only social media but also public pay great attention to them. The government is facing a severe challenge of food safety management. According to the features of food safety events such as short life cycle，regular development curve and so on，it is feasible to predict the heat trend of online public opinion in food safety area using Entropy Law and Markov Chain Model. Based on detailed analysis of hot food safety events through the amount of micro-blogs’and forums’posts as well as forums’comments and reposts，the weight of each indicator can be calculated. And then utilize Markov transfer matrix to predict the trend of online public opinion in food safety. After retrospective forecast of popular food safety events in 2013，the accuracy of prediction using Markov Chain Model is 85% means Markov Chain Model is suitable to forecast the trend of food safety online public opinions.