Abstract: One of the most important and promising branches of Artificial Intelligence (AI) is Machine Learning (ML),which strive to make a machine intelligent by “learning” from the data. Information Retrieval is also a popular and predominant technique having as one of its application, the ubiquitous Search Engine. Search Engine optimization (SEO) has seen remarkable advancements during the recent years. The objective of this paper is to optimize the existing predictive search mechanism by incorporating pattern based Machine Learning techniques, the association with a Semantic Database, Natural Language Processing of search query to produce more relevant predictions to the user. The main intention is to provide diversified but apt, intelligent predictions for both the diversified set of users whose domain of search queries is not constrained, as well as for the dedicated researchers whose domain will be confined, coupled with an optimal balance between the Response times, Relevancy of predictions.