<p>This study focuses on the four typical agrometeorological disasters in cold regions that frequently occur during the flowering period of spring maize in Northeast China, namely low temperature, drought, waterlogging, and insufficient sunlight. By using meteorological data, spring maize growth period, yield, and yield composition, and integrating statistical analysis and machine learning techniques, the Compound Stress Occurrence Index of low temperature, insufficient sunlight, drought and waterlogging during the Spring maize flowering reriod in Cold Regions (MCSI) was defined and constructed, using the bald tip rate as a characteristic parameter. The relationship among disaster-causing factors, yield composition, and yield were established hierarchically, the entire compound disaster process was analyzed, revealing the impact mechanism of the compound stress on maize yield. Relevant evaluation models were established and threshold indicators were divided. The results indicated that the critical threshold values of the compound disaster indicators of light, moderate and severe levels during the spring maize flowering period in cold regions are [0.31, 0.42), [0.42, 0.52), and [0.52, 0.76], respectively, and the yield loss rates are [0.4%, 3.5%), [3.5%, 7%), and [7%, 17.5%], respectively. The verification results of 446 samples showed that the average proportion of stations with consistent or a one-level difference in yield loss grades is 88.3%. It indicated that this index was capable of evaluating the yield loss caused by compound disasters relatively well. The research findings can enrich the monitoring, early warning and assessment approaches of agrometeorological disasters in the high latitudes at the technical level, providing meteorological references for ensuring the safe production of spring maize in Northeast China and regional disaster reduction and efficiency enhancement.</p>