The maintenance problem is of great importance in all engineering systems. In a vertically integrated electric industry, the system operator schedules both the generation and transmission maintenance. In a restructured environment, an independent entity called Independent System Operator (ISO) receives the generation and transmission owners’ preferred schedules and coordinates the final maintenance schedule as a system supervisor. In most models proposed for transmission maintenance scheduling, the adequate reliability of the system is considered as the most important constraint and then transmission maintenance is optimally scheduled. Since 2001, assessing the impact of the intentional attacks on the power systems has grown in importance. The classic criteria and power system reliability indices cannot take into account the intentional attacks on the power systems and their consequences, properly. In this thesis, at the first step, a new model is proposed to assess the vulnerability of a power system over a time horizon. The "time aspect" of the intentional attacks has not been investigated in the previous researches. The outcome of this step is a bilevel time-phased power system vulnerability assessment model. The Duality Theorem is used to convert the mentioned bilevel programming problem into a single level programming problem. At the second step, this model is used to formulate a new scheduling transmission lines maintenance model. In this new formulation, the maintenance scheduling problem is performed as a multilevel programming, while power system vulnerability as well as power system reliability are taken into account. The proposed models are applied to the standard Garver 6-Bus Test System and IEEE 24-Bus RTS and the results appreciate the capability of the proposed models. Keywords: Duality theorem, multilevel programming, power system vulnerability, transmission maintenance scheduling.