Solar energy has an important role to achieve the goal of replacing fossil fuels and significant potential to reduce greenhouse gas emissions. Accurate information on solar radiation is very essential for engineers, architects and agriculturist to design the energy systems based on the solar source. The sunshine duration and air temperature are measured by most of the meteorological services in all over the world but global solar radiation measurements are very rare and some of the data are missing. At this point, estimation of solar radiation where stations are not available plays an important role. Different models have been developed in the literature to estimate solar radiation. Angstrom-Prescott sunshine based model is widely used one and also there are some other approaches based on Angstrom model in the literature. In this study, linear, quadratic and cubic empirical as a general equation for throughout the year are generated to estimate global solar radiation in Eastern Mediterranean Region (EMR) which covers the four main cities (Adana, Mersin, Antakya and Kahramanmaras) by using the meteorological data in the Turkish State Meteorological Services. Regression models were estimated for each month separately and annually by curve estimation techniques with MINITAB statistical program. The monthly linear, quadratic and cubic models for estimating monthly average global solar radiation are validated as well. Finally, a comparison between monthly models and general models is performed by statistical test methods such as R-2, MPE and MAPE. According to statistical test results, the use of cubic general model for EMR is recommended. (C) 2014 Elsevier Ltd. All rights reserved.