Japanese logbook data have high spatial and temporal coverages in the Indian Ocean compare to those of observer data. However, the logbook data includes a large number of under-reporting catches for sharks, which makes it difficult to use directly the logbook data for the estimation of annual catch rates for sharks in the Indian Ocean. In order to solve the issue, the author used a statistical data method to filter Japanese logbook data and removed the systematic annual trends by reducing the set-by-set data with low reporting rates of catch for sharks using information on observer data. The reliability of the filtering method was validated using annual nominal CPUEs of tunas and sharks.