An effective method is proposed, in the present work, to reduce considerably the high errors normal in solar radiation data obtained by actinographs or by non-calibrated Eppley-type pyranometers. The method is based on constructing two annual distributions of clear sky solar radiation in each location studied. One of them should be constructed using measurements of an accurate instrument for this purpose, while the other is based on the actinograph or/and non-calibrated Eppley-type pyranometer under utilization. Then the correction factor can be deduced from these distributions The correction factor for a Robitzsch-type pyranometer has been found to be strongly dependent on the ambient temperature. Unlike the case of the Robitzsch instrument, the correction factor for Eppley instrument registrations is practically independent of weather conditions, but it depends highly on the length of the non-calibrated period. Validation of the proposed method is provided by comparison of measurements between those of an accurate instrument and the corrected Robitzsch measurements. it was found that the corrected values are reasonably accurate. A substantial reduction in the standard procedure registers is obtained.