The stability analysis considers the performance of generators and its ability to maintain synchronism under sudden changes in the power system. The response of these systems is non-linear, and its analysis must include deviations of rotor angles, rotor speeds, power flows in the system and voltages at machine's terminals. The use of computational tools is proposed in previous works, aiming for fast and effective solutions under high disturbances. Implicit integration and neural networks have been used to solve the system model, however simulation time does not comply with real time requirements. In this paper, explicit integration is used to solve the model's equations and the computing is accomplished using Graphic Processing Units (GPU) technology. The algorithm is implemented in MATLAB® and its main goal is to simulate the behavior of electrical power systems, looking to satisfy with real time requirements and to decrease simulation time using parallel processing technology.