In the current scenario, the social media demands the growing popularity of any services is entirely based on the huge number of user communications in form of customer comments, reviews and opinions. Therefore, it is highly needed that in any social networks, electronic media or in any services which resides in blogging the generated user communication should be of quality. The fake generation of user communications such as customer comments, reviews, and opinions could mislead the user, whereas the promotion of the product grows to be high. Sentiment Analysis (SA) is introduced to examine the opinions. Usually, sentiment analysis is defined as the area of study to examine people's emotion, reviews, and attitudes from written languages. The primary aim of this work is to develop an automated system, which would process customer fake review. Enhanced Random Forest (ERF) algorithm is proposed for detecting the reviews. The performance of ERF is analyzed by comparing it with the other ML algorithms such as decision tree, Naive Bayes, SVM, and traditional random forest algorithm. From the experimental results, it is noticed that the proposed ERF algorithm has higher performance ratio and minimum execution time than other ML algorithms.