Appropriate size and location of distributed generation (DG) play a significant role in minimizing power losses in distribution systems. This paper represents techniques to minimize power losses in a distribution feeder by optimizing DG model in terms of size, location and operating point of DG. Sensitivity analysis for power losses in terms of DG size and DG operating point has been performed. The proposed sensitivity indices can indicate the changes in power losses with respect to DG current injection. The proposed techniques have been developed with considering load characteristics and representing loads with constant impedance and constant current models, separately. The optimal size and location of DG in a distribution feeder can be obtained through the developed techniques, with minimum effort. The proposed techniques have been tested on a practical long radial system and results are reported. Test results have proven that up to eighty-six percent of real power loss can be reduced with a DG of optimal size, located at optimal place in the feeder.