This paper proposed a diagnostic supporting tool - i+DiaKAW (intelligent and interactive knowledge acquisition workbench), which automatically extracts useful knowledge from massive medical data by applying various data mining techniques for supporting real medical diagnosis. To fulfill this effort, i+DiaKAW has been developed to provide a novel dynamic on-line decision tree learning scheme - i+Learning (intelligent and incremental learning) methodology that is able to incrementally incorporate the new incoming data with an existing decision tree without relearning the entire tree from scratch. i+Learning strategy makes up for the traditional incremental decision tree learning algorithms by concerning the new available features in addition to the new coming instances. Such theory is a new attempt and is designed specifically for bridging the current gaps. The empirical results demonstrate a comparison of performance of various classification algorithms on several real-life datasets from UCI repository.