目前支持向量机回归算法只适用于单输入单输出系统,在处理复杂多变量系统时多采用构造一系列单变量支持向量机模型的方法,增加了模型与控制的复杂性。针对这一问题,提出了面向多输入多输出系统的支持向量机线性回归和非线性回归方法,两种方法的不同在于映射到高维空间后的运算方法不同。与单变量支持向量机回归建模相比,采用该方法建立的多变量系统模型结构较为简单,且在小样本数的情况下,模型训练速度较快。最后将该方法应用到甲基丙烯酸甲脂的间歇聚合反应过程中,仿真结果表明该方法的有效性。 At present, the support vector machine regression algorithm is only suitable for single-input and single-output systems. When dealing with complex multivariable systems, a series of univariate support vector machine models are used to increase the complexity of the model and control. To solve this problem, a linear regression and nonlinear regression method based on support vector machines (SVMs) for multi-input and multi-output systems are proposed. The difference between the two methods lies in the different operation methods after mapping to high-dimensional space. Compared with the univariate support vector regression modeling, the multivariable system model constructed by this method is relatively simple, and in the case of small sample size, the model training speed is faster. Finally, the method is applied to the batch polymerization of methyl methacrylate. The simulation results show the effectiveness of the method.