The method of bioelectrical impedance-based visceral fat estimation is in advance of other methods such as X-ray CT or MRI from the point views of cost and safety. However, it requires complex and diversity signal analysis and modeling to realize its high estimation accuracy. In response to this requirement, complication of feature attributes and simplification in selection of them to model the estimation has been proposed and evaluated in this paper. The complication of feature attributes is realized by applying priori knowledge and by employing the idea of cardinality as quantitative evaluation index. The simplification of the estimation model is realized by employing Akaike information criteria. The experiments were conducted to evaluate the proposed method and the results prove high estimation accuracy and stability of the proposed method.