The high toxicity and low biodegradability of the phenolic pollutants destroyed the balance of the environment and influenced human health seriously. Here, we developed a three-dimensional coloremetric sensor array for discriminating and determinating phenolic pollutants basing on the distinct Cu/nucleotides MOFs. Firstly, three laccase-mimic Cu/MOFs (Cu/AMP, Cu/CMP, and Cu/GMP) were obtained by regulating the molar ratio of Cu²⁺ and nucleotides. Then the Cu/MOFs as the recognition elements of the sensor array catalyzed the pollutants-4-AAP-H₂O₂ system, obtaining the colored benzoquinone products. Subsequently, the data array obtaining from the combined training matrix (3 Cu/MOFs × 6 pollutants × 5 replicates) was projected into a new dimensional space to obtain the 3D canonical scores, and classified into individual clusters by introducing LDA method. No overlap in their respective LDA plots for the six phenolic pollutants with different concentrations suggested the prominent discriminating performance of the sensor array. Furthermore, the sensor array exhibited high selectivity compared to the “lock-and-key” sensors even other active matrices coexisting in water samples. Importantly, the most influential discrimination factor was used to monitor the levels of the six targets, evidencing the potential application in assessing water pollution and maintaining human health.