Self-rectifying analog memristors have emerged as promising components for neuromorphic computing systems due to their inherent rectifying behavior and analog resistance states. Among these devices, BiFeO3 (BFO) memristors have shown exceptional performance, attributed to the accumulation and migration of oxygen vacancy (Vo··). However, the movement of Vo·· within the structure of the device presents challenges in optimizing their performance. To address this, the insertion of an interfacial layer has been proposed as a strategy to change the movement of Vo·· and enhance the behavior of memristor. In this study, we investigate the optimization of self-rectifying analog memristors by inserting an interfacial layer in BFO memristors. The more significant nonlinearity in high resistance state branch we observed in the current–voltage relationship leads to better rectifying behavior and a larger on/off ratio at room temperature, which indicates that the interfacial layer improves rectifying behavior. Moreover, we propose a model based on the modulation of the interfacial barrier to elucidate the impact of the interfacial layer on the BFO memristor. These findings provide insight into the design principles for optimizing self-rectifying analog memristors, with potential applications in neuromorphic computing.