Click-Through-Rate(CTR) prediction model based on deep neural network faces difficulties to deal with user behavior sequence effectively and efficiently while modeling preferences of users. This paper proposes a CTR prediction model named Long and Short Term Interest Network (LSTIN) to make full use of the context information and order information of user history records, in order to improve the accuracy and training efficiency of CTR prediction model. Based on attention mechanism Transformer and activation unit structure are used to model long-term and short-term user interests. The latter is processed by RNN and CNN further. Eventually, a fully-connected neural network is applied for prediction. Different from DeepFM and Deep-Interest-Network(DIN) in experiments on Amazon public dataset, LSTIN achieves a modeling with context and order information of user history and achieves distinguishing the long-term and short-term interests of users, which improves the performance and keeps training efficiency of CTR prediction model.