The extreme high temperature in western North America (WNA) exert profound impacts on industrial and agricultural production, public health, and trigger catastrophic wildfires. Exploring the underlying mechanisms influencing extreme high temperature days over WNA (WEHDs) and improving the seasonal prediction are of great scientific and social significance. This study reveals that two independent precursor signals, persistent negative sea surface temperature (SST) anomalies in tropical eastern Pacific and the cooling tendency in subtropical Atlantic SST exhibit significant influence on WEHDs. A physical-based empirical model constructed using these two predictors exhibits robust independent prediction skills. Guided by the underlying physical mechanisms, we integrate SST tendency fields as critical input features into convolutional neural network (CNN). The physically informed CNN achieves significantly improved performance and successfully forecasts the extreme WEHD events of 2021. The results emphasize the pivotal role of physical process understanding in advancing deep learning-based climate prediction.
 

extreme high temperature days,,western North America,CNN-based seasonal prediction