The objective of this paper is to extend the applicability of the GLR method to a wide range of practical systems. Most real systems are nonlinear, multivariate, and are best represented by input-output type of models. Kernel partial least squares (KPLS) models have been widely used to represent such systems. Therefore, in this paper, kernel PLS-based GLR method will be utilized in practice to improve damage detection in Structural Health Monitoring (SHM). The developed kernel PLS-based GLR technique combines the benefits of the multivariate input-output kernel PLS model and the statistical fault detection GLR statistic which showed performance in the cases where process models are not available. GLR is a well-known statistical detection method that relies on maximizing the detection probability for a given false alarm rate. To calculate the kernel PLS model, we use the data collected from the complex 3DOF spring-mass-dashpot system. The simulation results show improved performance of kernel PLS-based GLR in damage detection compared to the classical kernel PLS method.