The proposed model is a pair of stochastic processes, called state process and observation process. The random dynamic of meteorological regimes in random medium will be modellized by state process, an unobserved hidden homogeneous Markov chain. The observation process, which represents the clearness index sequence, will be formed by a stochastic differential equation. The parameters of the model are estimated from the real data using Maximum Likelihood (ML) estimation via Expectation Maximization (EM) algorithm. The simulated data will be used to predict the global solar radiation in the next time.