Improvement of the capability to cope with risk based on prior preventive management is very important for efficient operation of financial information system. In order to do this, understanding, analysis, and countermeasures for the risk that happened already in the past is essential. In this study, the defect data which happened in the financial information system including account system, business system and data feeding system during 4 years and 5 months were categorized and analyzed by the domain, defect factors, period, day of the week, phases of software development, and defect cause. As a result, it was identified that the defect data had characteristics and trends along the phase of software development, day of the week, and the cause, also that building risk prediction model was necessary for the risk management of whole financial domain due to the relation of the information systems.