Abstract: Near infrared spectroscopy is currently a highly versatile tool used in diverse fields. However, outlier samples strongly affect the performance of the prediction model in near infrared spectroscopy analysis. Therefore, detecting and eliminating the outlier samples is a major and important procedure in near infrared spectroscopy analysis. Using the outlier samples detection based on joint X-Y distances (ODXY) method, a special outlier samples detection method for multicomponent analysis was proposed and proved, termed as MODXY method. Experimental data was derived from the near infrared spectra of 80 corns. Based on these, the PLS models of moisture content, oil content, protein content and starch content were constructed by eliminating outlier samples using different outlier detection methods. The obtained models were compared in terms of performance by the predictive root mean square error (RMSEP) and the coefficient of determination ( R 2). The results showed that in most cases the MODXY method had better outlier sample recognition capability in NIR multicomponent analysis compared with other methods. Both ODXY method and MODXY method had their own suitable range, and they were more effective when the relative standard deviation of components was large enough.