Abstarct: Short time load forecasting has an important role in scheduling, utilizing and controlling of the power systems. accurate load forecasting results in practical decisions, such as maintenance, confident providence, unit commitment, economic dispatch and etc, to be improved. accurate short load forecasting according to the several effective load objects, such as weather conditions, daily, seasonal and temporary events with non linear load equations is a difficult job. solving above problems, here, short load forecasting performed by using artificial neural networks. artificial neural networks are able to obtain accurate non linear equations from input variables by using experienced data. here the point is to decrease the error in short load forecasting problem with considering special days by using an appropriate artificial neural network and also choosing most effective load objects. achieving this point feed forward artificial neural network with back propagation learning algorithm was studied and good results were obtained respectively. here to use neural network for the next 24 hour load forecasting, real load and weather data were obtained from Mashhad electricity center and also Mashhad weather forecasting center.