This paper focuses on product recommendation using user preference and online opinion mining. Multiple features of a product are analyzed to get user specific product. In this paper we have analyzed user review data of mobiles sold in online markets and then predicted the most suitable mobile for the user based on preference list provided by the user. To analyze the review, the textual data is processed in multiple stages like preprocessing, feature extraction, categorizing reviews based on features, and then predicting the polarity of the review data using several machines learning model like Logistic Regression, Support Vector Machines, Naive Bayes Classifier and Multi-Layer Perceptron Classifier. After this user entered preferences are analyzed to give user specific products.