Expanding target pesticide species and intelligent pesticide recognition were formidable challenges for existing cholinesterase inhibition methods. To improve this status, multi-active Mel-Cu nanozyme with mimetic Cu-N sites was prepared for the first time. It exhibited excellent laccase-like and peroxidase-like activities, and can respond to some pesticides beyond the detected range of enzyme inhibition methods, such as glyphosate, carbendazim, fumonisulfuron, etc., through coordination and hydrogen bonding. Inspired by the signal complementarity of Mel-Cu and cholinesterase, an integrated sensor array based on the Mel-Cu laccase-like activity, Mel-Cu peroxidase-like activity, acetylcholinesterase, and butyrylcholinesterase was creatively constructed. And it could successfully discriminate 12 pesticides at 0.5-50 μg/mL, which was significantly superior to traditional enzyme inhibition methods. Moreover, on the basis of above array, a unified stepwise prediction model was built using classification and regression algorithms in machine learning, which enabled concentration-independent qualitative identification as well as precise quantitative determination of multiple pesticide targets, simultaneously. The sensing accuracy was verified by blind sample analysis, in which the species was correctly identified and the concentration was predicted within 10% error, suggesting great intelligent recognition ability. Further, the proposed method also demonstrated significant immunity to interference and practical application feasibility, providing powerful means for pesticide residue analysis.