Medical informatics focuses upon the use of computer knowledge to various fields in medicine such as healthcare, medical research, and discerning medical learning. It solves various problems related to health environment as well as provides value-based health information to arrive at a decision. From the set of records, the optimized feature set is identified by applying Bee-based Harmony Search (BHS) algorithm and decision tree algorithm with a linear mathematical model. The BHS algorithm modifies the memory consideration rule of the Harmony Search. In Harmony Search, the exploitation is mainly controlled by HMCR value. Data analysis in healthcare is an exploring research area for building real-world decision support system. The developed predictive model can be effectively utilized for the existence of the type of diabetes by medical professionals for its evaluation. Regression analysis has been carried out for the determination of linearity among the predicted attributes PPG, A1c, and MBG level.