Tea quality assessment methods generally require elaborate sample preparation steps and often yield in consistent results. Over the last two decades, the diffuse reflectance spectroscopy (DRS) is emerging as a rapid and non-invasive tool for assessing quality parameters of several materials. In this study, we examined the DRS approach by creating a reflectance spectral imaging of 81 black crush, tear and curl (CTC) tea samples and eight tea quality parameters. Spectral models for each parameter were developed using the partial-least-squares regression (PLSR) approach. The ratio of performance deviation (RPD) was used to assess such spectral models. Results showed that moisture content, thearubigin components of TRSI and TRSII, total polyphenol contents and liquor brightness were accurately predicted with RPD values 3.63, 2.32, 2.24, 2.23 and 2.02, respectively. Prediction accuracies were moderate for thearubigin, total liquor color and theaflavin. The variable important projection (VIP) showed the wavelengths around near-infrared and shortwave-infrared regions strongly influence spectral reflectance of black tea. (C) 2016 Elsevier Ltd. All rights reserved.