Energy saving is extremely important in portable and hand-held devices like laptop and mobile phones. As applications become increasingly sophisticated and processing power increases, the most serious limitation on these devices is the available battery life. One promising technique that provides the best of both low-power and high-performance processors in the real time systems is Dynamic Voltage and Frequency Scaling (DVS). In this paper, we are working on the problem of power optimization for periodic independent real-time tasks that are scheduled over multiprocessors. Theoretically, Real Time tasks are specified with Worst Case Execution Time (WCET). During the operation of the system, the Actual Execution Time (AET) of a task could be a lot smaller than its WCET. We present an Adaptive real time DVS algorithm (Art-DVS) based on AET of tasks. It adapts two state-of-the-art scheduling algorithms Pre-DVS and Post-DVS. Simulation results show that the proposed Art-DVS algorithm performs satisfactorily on energy consumption in comparison to the previous approaches. © 2014 IEEE.