Distributed generation has become a popular research are with rapid increase on power demand. Optimal size and location of distributed generation are an important optimization problem to satisfy growing power demand. In this study, particle swarm optimization has been used for optimal placement and sizing of distributed generation to minimize power loss in distribution network considering different static load models such as constant power, constant current, constant impedance, residential, commercial, industrial and mixed load models. Results have shown that using this method, voltage profile of the distribution network has been improved and numbers of buses violating voltage limits have been reduced after distributed generation placement. Further power system operating point has been improved and loadability of network increased without violating bus voltage and line current limits. Analysis has been demonstrated on a 12.66 kV 33-bus radial distribution network. Mathematical models of the loads have been considered as voltage dependent exponential load model and practical values of the active and reactive power exponents for load models have been used to represent their influence on real power loss and voltage stability issues. Indices related with voltage stability and power loss presented in literature have been calculated to determine performance of the study. Besides less parameter requirement of particle swarm optimization method, it is easy to implement on large power systems and these characteristics make the method preferable.