The orderly development of marketization of power system brings new requirements for load forecasting in industrial park. This paper proposes a load forecasting method for industrial park based on deep learning and weight sharing. In forecasting models, the deep neural network is regarded as supervised-learning approach, weight sharing is deployed to analyze the correlation among various objectives, and the most related objective task is selected by load change rate. The validity of the algorithm is verified through the simulations performed by actual operating data from industrial park load system in Tianjin. The positive results demonstrate that the proposed algorithm can effectively improve the accuracy of load prediction and has high application value.