Blockage diagnosis is indispensable to ensure that microreactors operate effectively and stably. In this paper, we propose a data-based blockage diagnosis method using pressure sensors that can identify two blockage locations in a microreactor. The method identifies these locations by comparing measured pressure distribution data with prepared pressure distribution data for a case of blockage occurrence. The data are calculated by computational fluid dynamic (CFD) simulation. To avoid a combinatorial explosion of CFD simulations for database construction, the proposed method identifies blocked locations from just pressure distribution data when a single channel is blocked. The proposed method is applied to the blockage diagnosis problems of a microreactor with ten microchannels. The results of CFD simulations show that the method can accurately identify two blockage locations using fewer pressure sensors than there are microchannels.