Following the increase in of the information available on the Web, it is important to diversity of its users and the complexity of Web applications. One web application that has a diversity of users is a news website. Customizing a website with the characteristics of each user is called personalization. The purpose of this study is to study the methods used in giving news recommendations using user personalization. Collaborative filtering method (CF) is one method that groups users based on the nature of the user. This CF method can be applied using the k-nearest neighbor (KNN) algorithm. The proximity between users in this algorithm is sought using the Pearson correlation technique and cosine correlation. The best technique by considering the smallest value of prediction error evaluation will be applied to giving recommendations. Evaluation of these errors was tested by applying the formula Root Mean Square Error. The best evaluation results obtained in this study are the k-nearest neighbor algorithm with cosine correlation similiarity.