I propose to use the AHP (Analytic Hierarchy Process) method to determine the weights of environmental factors in a soil-landscape model used in knowledge-based automatic soil mapping. AHP provides a structured scheme to reduce the arbitrariness and improve the consistency in the weighting operation of a multivariate decision making process. In an AHP for soil mapping, the soil scientist first conducts a pair wise comparison between the considered environmental factors, using a nine-scale describing system ranging from “equally important as” to “extremely more important than”. Then, the computer performs a matrix computation to derive the weight for each environmental factor, using the input from the soil scientist. When a friendly user interface is provided, the soil scientist may feel that the AHP weighting process is simple and intuitive. An experiment in LaPointe Brook watershed, VT reveals that it is very easy for soil scientists to learn to use this method, and the soil scientist's knowledge of soil-landscape relationship can be more thoroughly and accurately represented using this method than using the limiting-factor (minimum) operation in the current SoLIM (Soil Land Inference Model) system.