Prediction of photovoltaic (PV) system generations is a powerful tool for managing the electric grids with multiple PV systems for reducing the instability of their electricity supply. For the prediction, weather and solar irradiance forecasting using a numerical meteorological model are performed in this study. The target country is Thailand which is in the tropical zone. In the tropics, cumulus and cumulonimbus appear frequently, and their behavior makes weather forecasting difficult. The correlations of the forecasted irradiances with the observations are about 0.8 or more in the intra-day, next-day, and 2-day ahead forecastings. It shows that the validity of solar irradiance forecasting in the tropics for the prediction of PV output. On cloudy days, the tendency of the fluctuation of solar irradiance is reproduced well in the forecasting, but a phase of the fluctuation is shifted in the forecasting.