<p>Soiling on solar photovoltaics causes substantial losses in transmittance and increases solar cell temperature, causing significant power losses. The present study investigated the impact of dust particles on solar cell performance by examining their thermophysical properties, such as density, size, thermal conductivity, and specific heat. The study employed machine learning techniques together with thermal analysis to predict solar cell temperature on soiled surfaces. ANSYS transient thermal analysis and Computational Fluid Dynamics simulations were conducted using the Shear Stress Transport k-ω turbulence model and discrete phase model to analyse the effects of varying dust thermophysical characteristics. Exploratory data analysis was used to identify the most significant variables for predictive modelling, utilising a random forest approach. The developed Random Forest model showed high predictive accuracy and achieved a coefficient of determination (<i>R</i>²) of 0.973 and a Mean Absolute Percentage Error of 0.18&#xa0;°C. Energy simulations revealed an average annual energy loss of 22.18% caused by soiling, of which about 4.5% was attributed to dust-induced thermal effects. This corresponded to an estimated annual financial loss of US$1 907.26 for a 25 kWp photovoltaic system. The results highlighted the critical need for effective soiling management to improve solar panel efficiency and economic returns. The study emphasised the importance of accurately predicting soiled solar cell temperatures and illustrated that random forests are a viable method for temperature prediction.</p>