In response to the severe threat of spring cold air to the survival rate of crayfish seedlings in rice-crayfish co-culture areas of the middle and lower reaches of the Yangtze River, this study aims to investigate the impact patterns across different stages of cold air, identify core disaster-inducing factors, and construct a prediction model. Based on 39 controlled trials of staged seedling stocking (greenhouse vs. open-air groups) during 8 cold air events from March to April (2022-2024) in Jingzhou, we systematically analyzed differences in crayfish survival rates under varying cold air intensities, environmental indicators, and stages. Key meteorological disaster indicators were screened, and a multiple regression model for 15-day survival rate prediction was established. Independent data from 2025 were used to validate model performance. The results indicated: (1) Survival rates exhibited a "high-low-high" stage pattern. The survival rate was the highest (83.6% -85.6%) 4 days before the arrival of cold air ( D-4 ) and 2-3 days after the end of cold air ( D2-D3 ), but significantly decreased to 66.9%-70.8% during rapid cooling period (D-1 to D1) (P<0.05). (2) Maximum 24-h water temperature drop (ΔTw₂₄) was the core stressor, showing the strongest negative correlation with survival rate (R=-0.752, P<0.01). ΔTw₂₄>7°C was the critical threshold for acute cold damage (survival rate 57.5%). (3) Water temperature fluctuation (Vw) within 1-2 days post-stocking had a negative impact (R=-0.639, P<0.01), 2.6 times stronger than the positive effect of average temperature, indicating stability over absolute value.(4) The integrated prediction model (Ŷ=91.536-3.959X₁+4.284 X₂+1.237 X₃, R²=0.858, SE=4.51%) was externally validated. After merging datasets, R² increased to 0.95 (MAPE=5.5%), confirming that multi-process data integration significantly enhances model interpretability and stability. This study systematically reveals the dynamic impact mechanism of spring cold air on the survival rate of crayfish seedling stocking, clarifies the quantitative relationship between temperature fluctuation characteristics and disaster-causing thresholds, and proposes a prediction model based on multi-stage meteorological-environmental coupling indicators. It provides a quantitative tool for precise aquaculture decision-making and has important theoretical and practical significance for formulating climate-adaptive aquaculture strategies in the rice-crayfish co-culture areas of the middle and lower reaches of the Yangtze River and for disaster prevention and mitigation in aquaculture under extreme weather.