Because there are great demands of distinguishing multiple chemically similar analytes, chemical sensors for multivariate analyses have been developed rapidly in the past few decades. However, designing luminescent discriminative sensors based on a monosystem has been a challenge until now. In this work, we first develop a triemitting luminescent discriminative platform named RGB@TLU-2 with three different emission centers: blue-emitting center (BDC-NH 2 ), green-emitting (Tb@ BDC-SO 3 − ), and red-emitting center (rhodamine B, RhB). The different luminescent mechanisms (ligand emission, LMET emission, guest emission) in these emission centers endow RGB@TLU-2 with high cross-reactivity, which is essential for discriminating applications. To balance the three luminescent centers, all variables in the synthesis process are optimized carefully. Surprisingly, the RGB@TLU-2 shows a variety of luminescent response patterns when immersed into 12 inorganic anions. Two unsupervised multidimensional analysis methods, (principal component analysis and hierarchical cluster analysis), are used to explore the relationship between these anions. On the basis of the luminescent response of analytes, 5 response modes are obtained and 12 inorganic anions are classified into 6 groups. The sensing mechanisms are discussed in detail. Detection limits of typical anions Cr 2 O 7 2− , PO 4 3− , ClO − , and NO 2 − are calculated as 2.895 × 10