Solar energy is one of the oldest renewable sources in the world. It is freely available and could be easily harnessed to reduce reliance on hydrocarbon-based energies. Data on solar radiation play a very important role in the design, sizing and performance of energy systems and renewable energies. However, these data are not always available, especially in isolated areas. As a result, predicting solar radiation values is often the only practical way to acquire this data. In fact, the measured sequences of radiation values are only available for few localities or regions of each country. Fuzzy logic is classified among the techniques of artificial intelligence, it is characterized by its simulation of human reasoning,
fuzzy logic has contributed in the development of several fields, In estimation, fuzzy logic is used to solve complex problems of the prediction .In this work, six solar radiation fuzzy models have been implemented to estimate daily global solar radiation on a horizontal surface based on two parameters : ration between duration sunshine (S) and daylight hours (S0) and the astronomical season (winter, spring, summer and autumn). Batna city is selected for this study. Its meteorological data sets of ten years (1996–2005) have been collected from the database HelioClim1 (HC1). The obtained results show that the new fuzzy models are adept to predicting the daily global solar radiation with a good measuring accuracy (the model 6 provided the best results MBE = 0.472, RMSE =1.750 and MAE = 1.395) in this city