Personalization of movie recommendations is a widely researched topic. Personalization is usually carried out using local resources that are available at one's disposal. This local resource presents a snapshot of user preference at a particular moment. It doesn't address the long term user preferences. These concerns can be addressed using resources available with the user. This paper proposes a model that taps the user browsing history with emphasis on smartphone browsing history to personalize movie recommendations. The browsing history and movie plot summaries are used to generate a similarity score. The obtained score is incorporated into a latent factor model that computes latent user and item features. This model enables prediction of user ratings under sparsity and cold-start scenarios using user browsing history and eventually fetches movies that are similar to the ones the user liked. © 2017 IEEE.