Near infrared (NIR) spectroscopy was investigated to predict trashcontent and classify types of ginned cotton by using a fiber-optic in diffusereflectance mode. Different spectra preprocessing methods were compared, andpartial least-squares (PLS) regression was established to predict the trashcontent of ginned cotton. Discriminant analysis (DA) was used to classify varioustypes of lint and content level of trash. The correlation coefficient r was 0.906for optimal PLS model using three factors based on first-order derivativespectra, and RMSEC and RMSEP was 0.440 and 0.823 respectively. To classify ginnedcotton with and without plant trash, the accuracy rate reached 95.4% using 15principal components (PCs) via DA, whereas the prediction accuracy rate was only80.9% for the classification of sample types due to containing foreign fiber, andthe classification result for the content level of trash in lint was not good forthe samples without any preprocessing. The result indicated that the NIR spectraof sample can be used to predict trash content in ginned cotton, which is oftendisturbed by type, content and distribution of foreign matters, and the accuracyof some prediction model is unsatisfactory. In order to improve the predictionaccuracy, some methods would be applied in future research, such as pretreatmentaccording to acquisition request of solid sample, or using transmission mode.Classification1)资助