Estimating global solar radiation (Rsg) is required under missing data conditions. It is an important factor to study the spatial and temporal distribution of energy parameters for different purposes. Despite being the primary energy source of the climate system, solar radiation is measured only at a limited number of weather stations. The main objective of this paper is to evaluate different models to estimate the Rsg in the region of Castilla-La Mancha (CL-M) from data obtained from termopluviometric weathers station. In addition, spatial and temporal Rsg will be mapped. In C-LM, historical monthly average data from temperature and rainfall series are available. For estimation of Rsg values, several approaches were applied: 1) lineal regression (RL); 2) artificial neural networks (RNA), and 3) the Hargreaves model. The goodness-of-fit of each model was computed for each model. The best models were those based on ANNs using maximum and minimum temperature values and extraterrestrial radiation (Ra) data. The Hargreaves model showed also a proper behaviour. The relative errors found for model validation ranged between 5.8% and 8.3% for ANN models and between 7.1% and 9.7% for the Hargreaves model. By using Rsg historical data which are developed through the application of RNAs from termopluviometric data, the Rsg that characterizes C-LM has been mapping.