This paper presents a new framework to study the generation capacity expansion in a multi-stage horizon in the presence of strategic generation companies (GENCOs). The proposed three-level model is a pool-based network-constrained electricity market that is presented under uncertainty in the predicted load demand modeled by the discrete Markov model. The first level includes decisions related to investment aimed to maximize the total profit of all GENCOs in the planning horizon, while the second level entails decisions related to investment aimed at maximizing the total profit of each GENCO. The third level consists of maximizing social welfare where the power market is cleared. The three-level optimization problem is converted to a one-level problem through an auxiliary mixed integer linear programming (MILP) using primal–dual transformation and Karush–Kuhn–Tucker (KKT) conditions. The efficiency of the proposed framework is examined on MAZANDARAN regional electric company (MREC) transmission network – a part of the Iranian interconnected power system. Simulation results confirm that the proposed framework could be a useful tool for analyzing the behaviour of investment in electricity markets in the presence of strategic GENCOs.