A new SVM (Support Vector Machine) classifier-combination model, based on Hierarchical Partition approach, for enterprise credit assessment is proposed in this paper. Enterprise credit assessment is essentially an ordinal classification (or ranking), and the popular multi-classification technique does not deal with the ordinal information in the sample data. Hierarchical Partition approach makes use of the ordinal characteristics to simplify the structure of classifiers, as well as to decrease the training and testing time. It can also be used to solve the problems where the imbalance sample distribution and different loss-costs among different ranks present. Experimental results show that the generalization ability of the new model using Hierarchical Partition approach are better than that of neural network model and traditional 1-vs-1 combined SVM model.