In this study, by using a data envelopment analysis (DEA) model we assess and compare the efficiency of some electricity distribution districts as decision-making units (DMUs) within the Tehran provincial electricity distribution company located in Tehran province, Iran. We termed the model as electric power distribution as the choice of inputs are based on a set of resources that supply and distribute electricity and the outputs will be such as the amount of electrical energy sold and the units of energy delivered. The definition of outputs in most studies is only limited to desirable outputs. However, the outputs in this study consist of both desirable and undesirable variables. Hence, we first extend the conventional DEA model to deal with undesirable outputs. We then provide the efficiency value of each district. Six out of twelve districts were efficient with efficiency score of 1, while others are inefficient with efficiency scores ranged from 0.5594 to 0.9738. The lesser efficiency value can be obtained due to losses in medium voltage transmission lines, low voltage transmission lines and transformers. To address the issues, some solutions such as modifying the network configuration using the existing methods for decreasing power losses and improving the voltage value in distribution network and dividing the power cable into several sections to switch to higher electrical power can be applied.