<p>Memristors have attracted a great deal of attention due to their potential for storage and computation. In this paper, we designed an Au/Al:HfO₂/ZnO/Al:HfO₂/FTO three-layer structure of the memristor. We doped the HfO₂ film was doped with different concentrations of Al elements to explore its resistive switching (RS) performance and photocatalytic synaptic properties. The experimental results showed that the device displayed a reliable switching ratio (~ 10²) under 100 consecutive cycles when the Al doping concentration is 10%. We further analyzed the conduction mechanism governing the device’s high-resistance state (HRS) and low-resistance states (LRS) under varying voltages. The device demonstrated similar resistive switching behavior across the scanning voltages, suggesting its ability to mimic the enhancing and inhibiting behaviors of biological synapses. Additionally, we explored the optoelectronic synaptic properties of the 10% Al:HfO₂/ZnO/10% Al:HfO₂ triple-structured memristor, successfully simulating the plasticity of biological neuronal synapses. Finally, we trained a convolutional neural network on the MNIST and Fashion_MNIST datasets, achieving recognition rates of 95.4% and 84.5%, respectively. These findings provide a unique opportunity for future advances in non-transitory memory and in neuromorphic computing.</p>