The collection and process in four diagnostic information of traditional Chinese medicine (TCM) are unsure and fuzzy. So it badly affects the veracity of diagnostic conclusion. This paper, based on former research work and eight principal theory of TCM, develops an intelligent ensemble TCM Eight Principal diagnosis system. Firstly, the acquisition of objective-oriented and quantitative four diagnostic information of TCM is an important prerequisite for intelligent diagnostic systems. In this paper, the system utilized tongue instrument, color inspection instrument, pulse diagnosis instrument, and point diagnosis instrument, to collect TCM symptoms and signs. And secondly, based on incremental symmetrical uncertainty in four views, the critical symptoms subsets are selected from original and noise symptoms. Then the syndrome type of a new case is predicated automatically by use of the intelligent ensemble TCMEPD system. In this paper, average 17.75 critical symptoms are selected from 177 symptoms and corresponding syndrome of 566 new cases are recognized.