The vast growth of information on the Internet as well as number of visitors to websites add some key challenges to recommender systems such as providing accurate estimation, handling many recommendations efficiently and coping with the vast growth of number of visitors in the system.Therefore, new recommender system technologies are needed that can quickly produce high quality recommendations even for huge data sets. The task of Collaborative filtering is to predict the utility of services to the active user based on the user's previous likings or database of user votes from a population of other users.The accuracy of user similarity is the key to the success of the recommendation for computing predictions. Recommender model predict user interest with the help of prediction function Prediction is a numerical value expressing the likelihood of the active user prefers the item. At various sites, a users' feedback is often continuously collected over a long period.By including some time function we modifiy prediction function to improve the efficiency of the recommender system.