Bisphenols, as common industrial raw materials, are widely used in food packaging such as plastics. However, their migration and residue may affect the hormone secretion of the human body and then lead to health problems. Therefore, a low-cost, rapid and simple detection method that can simultaneously detect multiple bisphenols is very necessary. In this work, two types of manganese single-atom nanozymes with excellent peroxidase-like activity were synthesized with graphyne as a support. A high-throughput colorimetric sensor array was constructed using three types of nanozymes (Mn-GY, Mn-GY-2N, GY-2N) to distinguish various bisphenols. Due to the absorption of bisphenol molecules on the surface of nanozymes, the activity of nanozymes decreases differently, when different bisphenols are added to the catalytic system. The results proved that the prepared sensor had good linear relationships at both low and high concentrations for determination of five bisphenols. The LODs of BPA, BPS, BPF, BPAF, and Diphenolic Acid were 0.443, 0.280, 0.277, 0.424, and 0.326 μM respectively. Compared with traditional sensors, the sensor array can simultaneously detect multiple analytes with high throughput, showing great advantage in dealing with complex samples. Combined with machine learning algorithms, five bisphenols can be successfully identified by the obtained array data. The sensor array also demonstrated excellent performance in the detection of both mixed samples and real samples. This high-throughput colorimetric sensor array achieves accurate and sensitive detection of bisphenol substances, providing new means and ideas for enhancing food safety. At the same time, the simple and rapid identification of structurally similar compounds demonstrates its potential for more precise analysis, providing possibilities for future development.