Transmission expansion planning (TEP) is critical for the power grid to meet fast growing demand in the future. Traditional TEP model does not utilize the flexibility in the transmission network that is considered as static assets. However, as the load profile may have different seasonal patterns, the optimal network configuration could be very different for different seasons in the planning horizon. Therefore, this paper proposes to incorporate seasonal network optimization (SNO) into the traditional TEP model. SNO dynamically optimizes the network for each season of each planning epoch. Two TEP-SNO models are proposed to investigate the benefits of optimizing the status of (i) existing branches, and (ii) existing and new branches, respectively. Numerical simulations demonstrate the effectiveness of the proposed TEP-SNO models. It is shown that SNO can improve system operational efficiency, defer investment of new transmission elements, and reduce the total cost.