For the data monitoring and fault early warning of new energy grid connection, the equipment status data, Supervisory Control And Data Acquisition (SCADA) system data, Wide Area Measurement System (WAMS) system data and other multi-source complex data were synchronized into the remote monitoring and early warning system of new energy station. The multi-source information fusion technology was used to analyze the data, and a new energy station operation parameter prediction method based on Multiple Extremum Learning Particle Swarm Optimization (MELPSO) algorithm was proposed. The correctness and effectiveness of the monitoring and early warning system was proved by comparing the predicted results of operational parameters with the measured results. The system can monitor the operation status of equipment in new energy station in real time, and has high prediction accuracy. It can judge the fault trend in advance and reduce the impact of new energy grid connection on the safe and stable operation of power grid.