This study introduces a new method for evaluating storm events which triggered slope failures, by using antecedent precipitation indices （APIs）. The method uses novel one-dimensional and two-dimensional diagrams, which examine excesses of APIs over their past maxima. While each of APIs with various half-life times（HLTs）is examined in the one-dimensional diagram, the two-dimensional diagram evaluates each pair of APIs with various HLTs. The method was applied to ５sediment disasters, and results indicated that the method could detect excesses of APIs over their past maxima at occurrences of slope failures for all disasters except for Maniwa disaster. Especially, the two-dimensional diagram was thought to be effective because it can detect anomaly in rainfall more credibly than the widely-used warning systems which use hourly rainfall, water storages in a tank model, and APIs with HLTs of 72 and 1.5h, as rainfall indices. For Maniwa disaster, some slope failures occurred before excesses of APIs were detected by the proposed method. This indicated increases in slope failure vulnerabilities in the region which suffered from treewindfall disasters.