Electric power network expanded rapidly in recent decades due of the excessive need of electricity in every aspect of life, including critical infrastructures such as medical services, and transportation and communication systems. Natural disasters are one of the major reasons of electricity outage. It is extremely important to restore electrical energy in the shortest time possible after a disaster. This paper proposes a decision support method for electric system operators to restore electricity to the critical loads in a distribution system after an earthquake. The proposed method employs Markov Decision Process to find the optimal restoration scheme based on the Probability of Failure of critical structures determined by using the Peak Ground Acceleration values recorded by observatories and earthquake research centers during earthquakes.