The support vector machine (SVM) method proposed by Vapnik [7] has been widely applied in the pattern recognition
and nonlinear regression problems. Additionally, SVM has found an increasing amount of applications in modeling and
control of many chemical processes [8-10], because of its good modeling performance. In contrast to NN, the primary
advantage of SVM is that, for a given data-based modeling problem with a finite set of samples, it can automatically
derive the optimal network structure in respect of generalization error [7-10]. In this sense, it can be expected, the SVM method can achieve a better prediction performance of the flooding velocity prediction with limited samples of the flooding data.