In this modern age of digitization, everyone uses online services for various day-to-day activities provided by numerous E- commerce websites. Based on the service provided we write reviews which explain and help other users to make decision about service available. These reviews or collection of data is very useful to improve and determine the pros and cons of service. Sentiment analysis is a field of Data Mining where user reviews are collected, analyzed and categorized in positive, negative and neutral reviews. In India, there has been a need for regional languages interfaces for better understanding about the web content and make it more user-friendly to native users. With that there was need for a system to categorize the user content of various regional languages. Here we propose a system to categorize movie reviews which are in Hindi into positive, negative and neutral for better decision making about the movie. The system utilizes a dictionary of positive and negative word list and phrases to determine the polarity of review. Special cases such as Negation handling, overall review of the movie is successfully implemented and tested on a dataset containing 500 reviews giving a total precision: 85.15% and recall: 97.23%.