Dust accumulation severely degrades photovoltaic performance in arid regions by reducing light transmittance and altering module thermal characteristics. However, quantitative understanding of the coupled effects of dust thickness and wind direction across their full parametric ranges remains limited, and the direct thermal impact of dust thickness on module temperature has not been systematically quantified through computational fluid dynamics simulation. This study employs computational fluid dynamics to systematically investigate the influence of dust thickness (0-10 mu m) and wind direction angle (0 degrees-180 degrees, at 15 degrees resolution) on module temperature. The results reveal a non-monotonic relationship between wind direction and temperature. Specifically, at 90 degrees angle of attack, minimizing turbulence intensity (62% lower than at 0 degrees) significantly inhibits convective heat dissipation. This leads to a peak average temperature of 341.28 K, representing a 7.92 K elevation above the 0 degrees. Turbulence intensity exhibits a stronger negative correlation with temperature (R =-0.975) than wind speed, indicating it is the dominant factor governing convective cooling. For dust effects, increasing thickness first reduces and then increases temperature, with a transition near 2 mu m. Guided by the parametric simulations, a hybrid regression model was developed by calibrating computational fluid dynamics correlations with observational data. This model predicts average module temperature as a function of wind speed, wind direction, irradiance, and dust thickness. It demonstrates robust agreement with measurements, achieving an R2 of 0.824, RMSE of 6.60 K and MAE of 5.27 K. This work provides an efficient predictive tool for optimizing thermal management and cleaning strategies in dust-prone power plants.