<p>Rutin (Rt) and quercetin (Qt), two prominent bioactive flavonoids, are widely present in functional foods and beverages, offering significant health-promoting properties. Accurate quantification of these compounds is essential for ensuring food quality and safety. In this study, an advanced dual-mode sensor based on a ternary biochar HB/ZIF-67/ZnIn₂S₄ nanocomposite was developed for the simultaneous and ultrasensitive detection of Rt and Qt. The sensor demonstrated exceptional performance, featuring a broad linear range (0.005–600 µM) and remarkably low detection limits (0.27 nM for Rt and 0.35 nM for Qt). Excellent repeatability and reproducibility were achieved, and real sample tests in functional food matrices yielded high recovery rates (Rt: 98.7%–104.2%, Qt: 97.0%–103.0%) with relative standard deviations (RSDs) below 3.36%. To further improve prediction accuracy, a machine learning-assisted approach using a backpropagation neural network (BPNN) was integrated for intelligent data processing. This work presents a robust, dual-mode strategy that combines advanced, nanocomposite-based sensing with machine learning, offering a powerful tool for the rapid, precise, and intelligent monitoring of flavonoids in food systems. The proposed method holds great potential for applications in food safety control, quality evaluation, and next-generation smart detection technologies.</p>