Nanofluidic memristors, which use ions in electrolyte solutions as carriers, have been developed rapidly and brought new opportunities for the development of neuromorphic devices. Utilizing the transport and accumulation of ions in nanochannels to process information is an endeavor to realize the nanofluidic memristor. In this study, we report a new nanofluidic memristor, which is a polydimethylsiloxane (PDMS)-glass chip with two platinum (Pt) electrodes and well-aligned multi-nanochannels within PDMS for ion enrichment and depletion. The device not only exhibits typical bipolar memristive behavior and ion current rectification (ICR) but also demonstrates excellent endurance, maintaining stable performance after 100 sweep cycles. We systematically investigate the key factors affecting ion transport behavior in this memristor. The results show that the ICR ratio of the current–voltage (I–V) hysteresis curves decreases with increasing scan rate and solution concentration. Zeta potential measurements are introduced to reveal that the PDMS surface carries more negative charges in higher pH solutions, resulting in more pronounced memristive and ICR effects. Furthermore, our memristor can simulate short-term synaptic plasticity, such as paired-pulse facilitation (PPF) and paired-pulse depression (PPD), with a relatively low energy consumption of 12 pJ per spike per channel. Potentially, the inherent accessibility and robustness of our nanofluidic memristors facilitate the optimization of device structure and performance. These important observations and investigations lay a foundation for advancing energy-saving and efficient neuromorphic computing.