Abstract: The temperature variability of South China in May is investigated, identifying precursor signals in sea surface temperatures (SST) and exploring the potential physical processes influencing these variations. A ridge regression prediction model has been developed. The analysis reveals that during years with anomalously high (low) temperature in May, there are observed anticyclonic (cyclonic) circulations over the Ural Mountains and East Asia, along with anomalous cyclonic (anticyclonic) circulations near Lake Baikal. These conditions weaken (strengthen) the East Asian meridional circulation, reducing (intensifying) cold air activity. Concurrently, the subtropical high abnormally extends westward (retreats eastward) in the South China region, while the southwest winds weaken (strengthen).The key precursor SST signals for temperature anomalies in May are identified, primarily from the North Atlantic tripole pattern in the preceding winter and the basin-wide variability pattern in the Indian Ocean. Among these, SST signal of the North Atlantic Ocean shows the strongest correlation. When the North Atlantic Ocean SST precursor signal is in a positive (or negative) phase, it influences the meridional circulation to weaken (or strengthen) and reduces (or intensifies) cold air activity through the Eurasian teleconnection wave train. Simultaneously, the subtropical high extends westward (or retreats eastward) in South China, resulting in higher (or lower) temperature.The multivariate ridge regression prediction model for temperature in May, developed using precursor signals from the preceding winter, demonstrates good fitting results and predictive capability for anomalous years. The model's performance is validated through various statistical tests, including mean squared error (MSE) and correlation coefficients, which demonstrate its robustness and accuracy in predicting temperature anomalies in May. Results indicate that the ridge regression model offers a significant advantage over traditional multiple linear regression models in this context. The model's predictive power is particularly remarkable in capturing the overall trends and variations of temperature in May, although it exhibits some limitations in predicting extreme values. The research provides valuable insights into the climate dynamics of South China and offers a reliable tool for enhancing the accuracy of short-term climate forecasts in the region, and underscores the importance of considering large-scale climate signals, such as the North Atlantic Tripole and Indian Ocean Basin-wide variability, in the development of predictive models for regional climate anomalies. By incorporating these signals, the model can better account for the complex interactions between different climate systems, leading to more accurate and reliable forecasts. This approach not only enhances our understanding of the factors influencing temperature in South China in May, but also provides a framework for future research and operational forecasting efforts aimed at mitigating impacts of climate variability.