This paper describes knowledge extraction process using decision tree technique that provides highly interpretable and a good accuracy in incomplete information system. In previous study, many real world data sets have incomplete information which attempt to impute some values or simply deleting directly the missing values. This incomplete information introduces uncertainty into decision modeling evaluation. We integrate expert knowledge and source of data to overcome the pitfall of the uncertainty with fuzzy representation. The degree of uncertainty of rank objects is measured during decision modeling for generating simple and comprehensible decision rule sets. Keyword: decision tree, classification, uncertainty.