A novel irradiance model is presented that simulates solar irradiance as a stochastic process with two states. The model contains, as a sub-model, the stochastic insolation model, which divides time into sunny periods and periods when the sun is hidden behind clouds, based on a homogeneous recurrent Markov process with two states. The irradiance model by itself calculates the irradiance depending on the state indicated by the sub-model. Methods are shown of how to adjust the model based on data easily available from meteorological stations.