Radio frequency power amplifiers operate most efficiently in the compression region and become less efficient as signal peak-to-average power ratio increases. This is due to the amplifier spending more time operating below peak power and, thus operating below its maximum efficiency. However, the use of the envelope tracking technique can dramatically increase the efficiency of RF transmitters. In this work, a Volterra series-based behavioral model with compressed sampling is developed for an efficiency optimized envelope tracking power amplifier. The proposed method reduces the total number of coefficients needed by the Volterra series model and enables accurate behavioral modeling of nonlinear envelope tracking power amplifiers. Experimental results carried on a GaN based power amplifier driven by a four-carrier 18 MHz WCDMA signal demonstrate the effectiveness of the proposed model and approach.