This research addresses microplastic contamination in water systems, where particles <5 mm are detected in 80% of global drinking water samples. We aim to develop a cost-effective, high-efficiency filtration system for scalable water treatment applications. This work introduces the first integrated system combining Surface Acoustic Wave (SAW) technology with machine learning algorithms (YOLOv8 and DeepSORT) for microplastic detection and filtration. The novel architecture features graphene-fabricated Interdigital Transducer (IDT) electrodes and replaces conventional lithium niobate with zinc oxide–glass lite substrate integrated with cellulose acetate butyrate, achieving 95% cost reduction while preserving performance. Current filtration technologies struggle to balance high removal efficiency with cost-effectiveness for large-scale deployment. Integrating acoustic manipulation with intelligent detection systems offers targeted, energy-efficient processing solutions. The SAW system uses 9 MHz signals transmitted to graphene-enhanced IDT electrodes, generating Rayleigh waves via the piezoelectric substrate. These waves concentrate microplastics at pressure antinodes, directing particles to a dedicated outlet while purified water exits independently. YOLOv8 and DeepSORT algorithms activate the signal generator only upon microplastic identification, optimizing energy consumption. The system achieves 93% removal efficiency, processing one liter every ten minutes. Machine learning detection shows 10–15% enhancement over baseline methods, identifying particles as small as 0.16 mm. Applications include residential filtration, aquaculture, and industrial processes. This integrated SAW-ML system provides scalable microplastic removal with substantial cost reduction, enabling widespread implementation. Future work includes ML hardware upgrades and real-time monitoring systems.