<p>Global warming has intensified heat waves in Mexico, increasing their frequency and severity. The predictions of the maximum air temperature (AT) for important heat wave occurrences by machine learning is discussed in this study. Monthly data, during summer from 2014-2024, downloaded from NASA's Giovanni and ERA5 include thirteen parameters. Events were rated according to the coping range using sensitivity analysis and the Pearson's correlation to identify the dominant inputs. The relationship of AT to land surface temperature (LST), black carbon (BC), aerosol optical depth (AOD), and carbon monoxide (CO) produced a strong connection. Three different machine learning models; random forest (RF), support vector regression (SVR), and multiple linear regression (MLR) were reviewed using three-fold cross-validation by randomly selecting one training (70 aspects), testing (20 aspects) and validate (10 aspects) for a single random forecast. The maximum temperature for each month based on AT data surpassed the criteria established by the IMD for heat wave readings. The evaluation of the machine learning models was done based on the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE) values. It was found that RF had the highest values R² (up to 0.98) and SVR had consistently the lowest RMSE or MAE making SVR the most reliable model to depend on. These findings display how machine learning methods can facilitate AT predictions of heat wave conditions to improve definitions of heat waves to base jurisdictional decisions on an eventual agricultural practices, urban planning, and climate modelling.</p>