Tinnitus problems affect a significant portion of the population and are difficult to treat. Treatment processes are plentiful, yet not completely understood. In this dissertation, we present a knowledge discovery approach which can be used to build a decision support system for supporting tinnitus treatment. Our approach is based on a significant enlargement of the initial tinnitus database by adding many new tables containing new temporal features related to tinnitus evaluation and treatment outcome. Research presented in this thesis includes knowledge discovery with temporal, text, and quantitative data from a patient dataset of 3013 visits representing 758 unique patient tuples. Additionally, a new rule generating technique and clustering methods are presented and used to develop additional new temporal features and knowledge in this complex domain. Of particular interest is the role that emotions play in treatment success for tinnitus following the TRT method developed by Dr. Pawel Jastreboff. The ultimate goal of understanding the relationships among the treatment factors and measurements in order to better understand tinnitus treatment will result in the design foundations of a decision support system to aid in tinnitus treatment effectiveness.