Credit risk has always been a major concern for banks and other financial intermediaries. The uncertain domain of risk assessment has long been in need of a reliable and consistent system to help simplify the decision making process. In recent years fuzzy logic technique has been in its wide-ranging use in modeling of uncertainties, vagueness, impreciseness and the human thought process. This paper attempts to develop a credit risk analysis system based on the fuzzy logic using two structures: traditional fuzzy reasoning of all inputs that map to one single output and stage-wise fuzzy reasoning of input parameters in accordance with their importance. The analysis of case studies shows the consistency and effectiveness of the second approach in making correct decision.