It became possible to simulate a qualitative and quantitative hybrid model by dividing ranges of the model's quantitative variables with ‘landmarks’ and by associating divided ranges and kinds of qualitative values with each other. When landmarks are uncertain, it is difficult to set appropriate landmarks to guarantee the accuracy of the simulation. In order to guarantee the accuracy of the simulation, the proposed method interpolates uncertain landmarks based on initial values and reliable landmarks of the hybrid model. First, candidate positions, where landmarks are interpolated, are decided by associating the range including an initial value and an initial qualitative value. Next, values of landmarks are estimated by applying monotone cubic spline interpolation to the reliable landmarks. Through the application of the proposed method to a hybrid model, the difference between the simulation result on the hybrid model and the result derived from the quantitative model is decreased by 0.17 compared to the method without the interpolation. © 2012 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.