In most cases, there is a substantial lack of weather data for renewable energy feasibility simulation. In this reason, generating weather data from limited monthly average information is essential in an implementation and simulation of smart grid system with a renewable energy. To predict solar radiation sequence and reduce the estimated error of the solar radiation in smart grid simulation, a novel solar data generating scheme which is called hybrid method of Markov transition matrices (MTM) and autoregressive model is developed. For case study to prove excellence of proposed hybrid method, an optimal MTM to estimate the daily solar radiation of Singapore is obtained by exploiting a historical data based on daily global solar radiation. Simulation results show that the root mean square error (RMSE) of proposed scheme is improved by approximately 50% comparing to that of the conventional MTM scheme.