Motion estimation is a key issue in any video compression algorithm. The main motive behind motion estimation algorithms is to optimize the motion estimation time or bit rate without compromising on PSNR value given by full search motion estimation algorithm i.e. benchmark algorithm. Real time videos are having varied and complex types of motion. Thus, priory information of video contents may help in building the robust and optimized motion estimation algorithm. Possibility of trapping the search in local minimum is addressed with the relevant parameter. Proposed algorithm is empowered by four aspects such as initial search point prediction, homogeneity parameters which give contents at macroblock level, global minimum predictor indicates closeness to near accurate minimum and adaptive search pattern selection based on prior information of homogeneity. Proposed algorithm is an incremental work to the algorithm by Humaira Nisar, Aamir Saeed Malik and Tae-Sun Choi on content adaptiveness in video in 2012. However, certain aspects like adaptive search pattern selection property and thresholds used in this work differ from original work. This algorithm is tested on JM18.4 of H.264/AVC and proving that the adaptiveness in search pattern selection helps in reduction of bit rate with maximum 4 % increase in motion estimation time compared to refered algorithm. This reduction in bit rate is not much hampering the encoding time and PSNR value. On the other hand this algorithm is performing better in terms of total encoding time for head and shoulder type of video sequences.