The resource leveling problem in Network planning is a combinatorial NP problem. In this present study in order to give the neural network description of resource leveling problem and to make the output correspondent with neurons, a new concept of augmented permute matrix is proposed. Some novel technologies are used when setting the energy function under time and resource constrains. An Embedded Hybrid Model combining Discrete-time Hopfield model and SA (DHNN-SA) is put forward to improve the optimization in which Hopfield servers as State Generator for the SA. At last, the simulating program built on DHNN-SA is created. The results after comparing with professional project management software show the energy function and hybrid model given in this study are highly efficient in solving resource leveling problem to some extent.