The effective and precise processing of visual information by the human eye primarily relies on the diverse contrasting functions achieved through synaptic regulation of ion transport in the retina. Developing a bio-inspired retina that uses ions as information carriers can more accurately replicate retina's natural signal processing capabilities, enabling high-performance machine vision. Herein, an ion-confined transport strategy is proposed to construct a bio-inspired retina by developing artificial synapses with inhibitory and excitatory contrasting functions. By fine-tuning the ionic hydrogel structures to control ion transport across the heterogeneous interfaces, inhibitory and excitatory synapses are realized to negatively or positively modulate the optical signal. The integration of these synapses facilitates advanced tasks such as image recognition and motion analysis. Moreover, as a proof of concept, guiding robot vehicles to perform path planning is demonstrated. This work offers a new idea for constructing the bio-inspired retina by precisely regulating ion transport, allowing it to reach a level closer to the biological retina.