Motion estimation is critical in motion coding. However, the fixed pattern of search center decision method in HEVC is lack of precision. Taking advantage of the surrounding coding unit information, a universal motion vector prediction framework is presented to improve the coding efficiency in HEVC. The proposed framework is composed of two parts: motion vector prediction (MVP) and search range (SR) selection. Firstly, a novel motion vector prediction (NMVP) method is presented to improve the coding efficiency. Secondly, an adaptive search range selection (ASRS) method is developed to reduce the coding complexity in motion estimation. The simulation results demonstrate that the overall bitrate can be reduced by 5.49% on average, up to 9.18% compared with HEVC reference software.