Video compression is required in applications like video network communications, video conference, broadcasting, live streaming and video storage. H.265/HEVC is the latest video compression standard, jointly developed by JCT-VC that provides the highest compression efficiency without significant loss in original video source quality. Among all the tools in HEVC encoder, Motion Estimation (ME) is one of the most complex tasks. The present paper analyses the ME algorithm present in HEVC standard reference software and proposes two improvements to the algorithm. Our results show a decrease on the computational complexity by almost 30% with negligible loss in the video quality.