Most of the DPSs have a strongly nonlinear spatiotemporal nature, and are featured by unknown parameter, boundary and even structure. To model these processes, the commonly used models often have a high-order, which makes them difficultly used for prediction. Based on the former spatiotemporal LS-SVM model, an improved online low-order spatiotemporal method is developed for the modeling of unknown and nonlinear DPSs. As we know, the information of a certain sensor may be represented by information of its neighboring sensors due to the spatial correlation between them. Making use of this feature, a kernel-space based spatial correlation analysis method is developed for deleting redundant SKFs, from which a low-order model can be achieved without loss of spatial information. On this basis, a LS-SVM model is constructed to represent the nonlinear temporal dynamics. Integration of the without-redundant SKFs and the LS-SVM temporal model, a low-order spatiotemporal model is obtained. Additional analysis and proof show that: (1) the proposed method has almost the same performance with the without-order-reduction spatiotemporal model; and (2) it has better modeling performance than the model with the same order achieved by the without-order-reduction one.