Classifying bug severity using dictionary based approach Bug tracking system allows user to report bugs that they encounter while operating the software. These bugs are then received by developers and they resolve these bugs according to their severity level. This task of assigning severity level is manual task that need expertise of assessing the severity level of reported bug. But if these reported bugs are large in number then manual process to assess severity level of bugs becomes very hectic. So there should be an automatic process to classify it so that bugs that need immediate fixation get resolved early. A few attempts have been made by researchers to automate the task. The approach followed in this paper has made an attempt to automate the bug severity classification using text mining technique and Naïve Bayes Multinomial classifier. This paper further proposes an approach of making this task more efficient using dictionary of bug terms.