Aiming at the massive, multi-source and complex battlefield data produced by joint operations under the condition of informationization, a Hadoop distributed data processing platform is proposed. Massive data are collected to analyze the elements of battle field situation (BS), and the particle swarm optimization (PSO) is used to optimize the extreme learning machines (ELM). The method of ELM is used to train the historical data of battlefield situation and construct the prediction model of battlefield situation, and Matlab 2018 is used to simulate the battlefield situation. The simulation results show that Hadoop is more efficient in processing massive battlefield data, and can effectively improve the prediction accuracy of battlefield situation, which provides a new method and way for assisting commanders to quickly grasp the complex battlefield situation.