ObjectiveUsing near-infrared spectroscopy and chemometrics, to establish a method for determination of multiple indicators in Eucommia ulmoides (EU) to achieve its rapid quality evaluation. Method83 batches of EU decoction slices were collected, and the content of pinoresinol diglucoside, chlorogenic acid and geniposidic acid in them was determined by high-performance liquid chromatography, the content of total flavonoids was determined by UV–Vis spectrophotometry, while the antioxidant activity of EU was measured by methods of DPPH, ABTS and FRAP, respectively. The near-infrared spectra of EU were acquired in diffuse reflectance mode, which were subjected to multivariate statistical analysis in relation with the above measured values. In order to achieve good model performance, various methods for spectra pretreatment and wavelength selection were explored. Especially, with the help of four wavelength selection methods based on swarm intelligence optimization, such as genetic algorithm (GA), gray wolf algorithm (GWO), improved global flower pollination algorithm (mgFPA), and improved sine and cosine optimization algorithm (ISCA), the partial least squares regression models were improved prominently. The model quality was evaluated by calculating the coefficient of determination (R2), root mean square error (RMSE) and residual prediction deviation (RPD) in the sets of calibration and prediction. ResultsRegardless of which indicator in EU, the model performance was fairly good, especially after the wavelength selection by GA and other three methods, where R2 > 0.9, RPD > 3 in both calibration and prediction, indicating that the models had good prediction accuracy. Based on the overall algorithmic efficiency and model results, we have the following ranking: GA > GWO > mgFPA > ISCA. ConclusionThe present models provide a novel method for rapid and comprehensive quality evaluation of EU, which may be of great significance for quality evaluation and quality control of EU.