The importance of plant alarm systems has grown with increases in the scale of plants and the sophistication of plant monitoring and control systems. In this paper, a two-layer cause-effect model is proposed that represents the cause-and-effect relationship between state variables. Using the proposed model, a derivation algorithm of sets of pairs of alarm variables and signs is proposed. The signs indicate the upper or lower limits of alarm variables. The derived sets of pairs are theoretically guaranteed to be able to qualitatively identify all assumed malfunctions. The proposed method using the two-layer cause-effect model was applied to a simple process, and the simulation results illustrated the usefulness of the method.