Memristors have been increasingly valued by researchers nowadays, and using the unique properties of memristors to build brain-like large-scale integrated circuits has great potential in the development direction of high energy efficiency peak neural networks in the future. Simulating brain function has far-reaching implications for the development of computers. This paper describes a synaptic circuit based on a memristor and amplifier to simulate axonal mounds’ action potential. A Cadence simulator was used to simulate synaptic current flowing out from dendrites in axonal colliculus and the relationship between the membrane voltage and the output voltage with time. This circuit can greatly reduce the mismatch of the circuit and has good matching characteristics. It also reduces the area of the circuit.