Wind farms begin to influence the power systems with the increasing amount of wind power penetration. The study of such influence justifies the need of a dynamic wind farm model comprising of a large number of generators, but detail models require high simulation computation time. An aggregation technique is required to reduce the model order while maintaining its accuracy. In this paper, a complete and an aggregated wind farm model with Doubly Fed Induction Generator (DFIG) are presented. Simulations have been carried out for both models and compared to demonstrate effectiveness of the aggregated model in terms of accuracy in approximation of the dynamic collective responses at the Point of Common Connection (PCC) and reduction in simulation computation time.