Average-value modeling (AVM) plays an important role in design and analysis of power-electronic based systems such as flyback converters. The existing AVM techniques such as state-space and circuit averaging have been developed by many researchers. In this paper, a new AVM of the second order flyback converter is introduced using the state space averaged value modeling approach. Here, the accuracy of the proposed AVM is enhanced by considering system basic parasitic losses. The proposed AVM is compared against the detailed simulation and previously-developed state-space model in large signal time-domain studies, and it is demonstrated to be accurate in predicting the original switching behavior.