Molten steel temperature is one of the parameters in converter end-point control. Accurate prediction of molten steel temperature has important guiding significance for the end-point control. However, most of the previous end-point prediction models are static models, which can only predict the end-point molten steel temperature in converter and can’t achieve dynamic prediction, and lead to the limited role of these models. To solve this problem, a data-driven prediction model of molten steel temperature in second-blowing stage in converter is proposed. Firstly, the model retrieves the similar cases in the historical case base through the process parameters in the main blowing stage of the new case based on the case-based reasoning algorithm. Secondly, the process parameters in the second blowing stage of the similar cases are used to train the relationship between the process parameters and the molten steel temperature based on the LSTM algorithm. Thirdly, the trained LSTM model is used to dynamically calculate the molten steel temperature in the second blowing stage of the new case. Finally, the model is verified by the actual production data of a steel plant, the results show that: the hit rate of prediction errors of the established model within the range of [﹣5℃,5℃], [﹣10℃,10℃] and [﹣15℃,15℃] are 39%, 73% and 90% respectively, which is higher than other dynamic prediction models. The model has a higher guiding role in practical production.