Distributed energy resources (DERs) planning and operation are not only the problems related to distribution system operator (DSO). Opeartion of DERs of distribution system also affects the operation of transmission systems. Both transmission system operator (TSO) and DSO can use flexibility provided by DERs. DERs can help to TSO to efficiently adresse transmission system problems such as power balance, voltage control, and congestion management, while DSO can use them for management of its local distribution system. TSO and DSO can operate in non-coordinated or coordinated manners. In the non-coordinated operation approach, transmission and distribution systems are operated in an independent manner without any interaction between TSO and DSOs. Thus, any operation scheme in one system is non-negotiable for the system operator on the other side which may lead to an undesirable and unacceptable operation due to power mismatch at the boundary buses. Even if they solve their sub-problems for predetermined boundary values, the obtained solution is not optimal. This means that the flexible resources in different systems cannot be efficiently exploited. Thus, the development of an interactive and coordinated operation framework for TSO and DSOs can considerably overcome the challenges and exploit the opportunities of high penetration DERs. In this dissertation, first, the concept of transmission and distribution systems coordination is introduced and developed to more and better exploit DERs. Then, two coordinated TSO-DSOs frameworks are proposed. The first framework is based on a cooperative perspective in which TSO and DSOs operate to optimize overall performance of the system. To implement this framework in a decentralized manner, an accelerated augmented Lagrangian (AAL)-based decentralized algorithm is proposed in which TSO and DSOs exchange a limited data to realize the cooperative operation. From the implementation perspective, the proposed decentralized algorithm has the advanyages of low computational and commucational burdun, and preserving privacy of independent system operators. Furthetmore, the cooperative framework can improve many power system technical indices such as voltage, congestion, and power losses, while reducing total operation cost of the system. In the second proosed framework which called non-cooperative, while TSO and DSOs are self-interested and individually solve thier sub-problems according to their own objectives and constraints, they are interactive and receive feedback from each other to understand the mutual effects of their decisions and give an appropriate response. This framework is implemented by a new decentralized algorithm based on the decomposition of boundary bus variables between transmission and distribution systems. In addition to the first decentralized algorithm’s advantages, this decemtralized algorithm has the advantage of very fast convergence rate. Using the non-cooperate frameworks, DSOs which may not necessarily attain their technical and economical objectives can selfishly optimez their objectives and attin better solutions rerspect to the cooperative framework. The selfish behavior of DSOs can also help to maximization of overall system indices, although not completely in line with it. The proposed coordinated TSO-DSOs frameworks are studied on various power system operation and planning problems including day ahead scheduling (DAS), joint active and reactive powers operation (JARPO), Volt-VAr optimization (VVO), and distributed generation planning (DGP) and results are compared with the non-coordinated framework. Furthermore, TSO’s and DSOs’ subproblems are formulated based on new linearized and convex AC optimal power flow (AC-OPF) models. The proposed frameworks have been studied on three integrated test power systems each of them consists of one transmission and some distribution systems. Simulation results con?rm the ef?ciency and effectivity of the proposed frameworks in terms of economic bene?ts, technical aspects such as power losses, and congestion management compared with the non-coordinated operation of transmission and distribution systems. If compared with a centralized approach and other decentralized methods, computational advantages are also con?rmed, such as achieving the optimal solution in cooperative framework and obtaining the Nash equiblirium in non-cooperative one with reasonable accuracy and time