This journal describes longest common subsequence algorithm used to classify the user navigation pattern for predicting the recommendation set for the online users and the proposed algorithm provides an efficient way for website improvement to the organizations according the sequence found by the proposed algorithm. That can make the website of any organization more efficient (for site improvement) and user friendly. The quality of the recommendations is measured by the two parameters that are accuracy, coverage and the Figure 10 shows that as the length of subsequence increases, the requirement of rearrangement of pages (site improvement) also increases.