FT-NIR and chemometrics are vital analytical tools for medicinal plant quality control. This study aimed to establish a rapid technique for quantifying sennoside A and B in Cassia plants (i.e., leaves, pods, and flowers) using Fourier Transformed Near-infrared Spectroscopy (FT-NIR). The calibration models were established by partial least squares (PLS), GA-PLS (Genetic Algorithm PLS), Bi-PLS (Backward Interval PLS), and Si-PLS (Synergy Interval PLS). Sennosides contents in Cassia were quantified using high-performance liquid chromatography. The outcomes indicated that GA-PLS and Si-PLS had the highest correlation coefficient of prediction (R2p) and lowest root-mean-square error of prediction (RMSEP). Then, the best prediction ability was achieved by Si-PLS (SA: R2p = 0.9732, RMSEP = 0.9720, 4400.761-4597.464 cm-1, 5203.003-5399.707 cm-1, and 5804.685-6001.388 cm-1; SB: R2p = 0.9924, RMSEP = 0.1220, 4547.3224-5091.152 cm-1, 5642.693-6186.521 cm-1 and 7285.747-7825.718 cm-1. The external validation of Si-PLS showed excellent results (R2p > 0.95), offering valuable theoretical guidance for developing an intelligent system to evaluate the quality of Cassia species in real-time.