In industrial production, it is necessary to accurately predict the load changes of multiple nodes before performing accurate load control. To solve this problem, this paper proposes a multinodal short-term load forecasting model based on participation factor of each node and the long-short-term memory (LSTM) network based ensemble learning. A hybrid ensemble strategy based on bootstrap sampling and weighted average sum is proposed to extract the deep features of multinodal load data, while the participation factor is adopted to represent the coupling between master node and slave nodes. Experimental results show the high accuracy of the proposed method in multinodal load forecasting.