Short-term power output forecasting of photovoltaic (PV) systems has become increasingly important to balance the supply and demand of grid electricity in an efficient manner. It is well known that the operating temperature has a strong influence on the power output of photovoltaic devices; therefore, improvements in simulating the short-term PV module temperature add to the accuracy of PV power output forecasting. In this study, a new empirical model is proposed, which is suitable for transient weather conditions. This model uses four variables to predict the temperature of PV modules, namely ambient temperature, irradiance, wind speed, and relative humidity. The model is applied and evaluated for various silicon-based PV technologies at two Southeast Asian locations (Singapore and Jayapura) on a 1-min basis. Results show root-mean-squared errors (RMSE) of 1.5-3.2 K, which marks an average improvement in RMSE of 46%, compared with similar existing empirical models.