Abstract: Persistent heatwaves severely impact economic and social activities; thus, it is crucial to study their causes and predictability. Using daily maximum temperature data from 142 stations in Hebei Province for July, along with NCEP/NCAR global daily and monthly reanalysis data from 1961 to 2023, this study analyzed the causes of heatwaves in July in Hebei. A heatwave forecasting model was developed using a deep learning algorithm LSTM (Long Short Term Memory networks). The results indicated that the average number of heatwave days in July 2023 in Hebei was 7.8 d more than normal. There were three prolonged heatwave events, with the strongest occurring from 5 July to 11 July. The anomalous westward and northward expansion and the large area of the western Pacific subtropical high, along with the persistent eastward and northward extension of the South Asian high, served as the background circulation patterns contributing to the ongoing development of this heatwave. Diabatic heating accompanied by strong descending motion was the direct cause of the hot and dry conditions. LSTM was employed to forecast conditions at four stations in central and southern Hebei Province, and the model predictions were all statistically significant. The assessment results showed that the AUC (Area Under Curve) values ranged from 0.55 to 0.65, with Baoding demonstrating the highest forecasting accuracy, indicating that the LSTM model has a moderate ability to predict daily maximum temperatures. Further analysis revealed that LSTM can effectively capture the heatwave trends at each station. However, significant discrepancies existed in the predicted start and end dates of the heatwaves and in the maximum air temperature values.