Abstract: The real-time load capacity assessment of transformers provides access to fully utilization of the potential load capacity of transformers while insuring the safe performance, which becomes the tendency of assets lean management. In this paper, the multi-thermal sources and heat transfer process of transformers were analyzed, and the absorbed power from sunshine radiation was introduced into the computation of winding hot-spot temperature. Then considering the DC resistance variation of winding and oil viscosity with oil temperature, the algorithms of thermal source power and thermal resistance were modified to improve the calculation accuracy of hot-spot temperature. Combined with the constraints of transformer load capacity, such as the threshold of top-oil temperature, hot-spot temperature, relative loss of life and the capacity of auxiliary equipment, a model of load capacity assessment for oil-immersed transformer was proposed, which takes different load types of transformers into account. The model was verified with the measured data of two transformers. The results show that the relative error of the improved algorithm of hot-spot temperature has been reduced by 2.5% compared with the IEC guide method. Besides, the results of load capacity assessment suggest that the actual load capacity of the test transformer is about 115% of the rated value, which means extra load capacity could be still exploited in the operation condition.