Traditional computer processing requires a large amount of data transmission between the processor and the storage unit, which limits the metering efficiency and the scalability of the architecture. To solve the limit, the concept of a brain-like parallel computing system based on artificial synapses was proposed. This work presents an optoelectronic synapse based on perovskite. The energy band of the perovskite material matches that of the metal oxide semiconductor, allowing the device to possess synaptic plasticity and thus be used to simulate biological synapses. Finally, based on the proof of concept, the neural network classification process using artificial synaptic devices was simulated.