based on conventional image sensors and data processing systems faces severe challenges in terms of efficiency and power consumption due to the physically separated memory and central processing unit (CPU), which are particularly important in the era of artificial intelligence (AI). [4,6] Fortunately, optoelectronic synapses are neuromorphic computing devices that rationally incorporate optical sensing, processing, and memory functions, [7][8][9][10] thereby enabling artificial visual systems to be realized with low power consumption and low mutual interference. [11][12][13] Optoelectronic synaptic transistors are one type of optoelectronic synapses that are widely applied in artificial visual systems due to their linear synaptic plasticity, configurable electrical modulation, and nondestructive operations. [14] Meanwhile, their multi-terminal device structure naturally permits the construction of large-area, high-density synaptic transistor arrays, capable of precise pixel addressing without sneak-path leakages for image perception and processing. [8,13,15] Despite these advantages, optoelectronic synaptic transistors with nanomaterials and/or nanostructured channels are often reliant on vacuum-based top-down semiconductor fabrication techniques, which are characterized by complicated processes, low efficiency, and high cost. [15,16] Printable oxide transistors have emerged as important optoelectronic synaptic devices to emulate the human visual cognitive system by integrating both photo-sensing and visual perception functions. However, the inferior electrical performance and low pixel density of printed oxide synaptic transistors impede their implementation of complex neuromorphic computing. Herein, a printable all-oxide optoelectronic synaptic transistor array based on a modified coffee-ring structure of indium tin oxide (ITO) is reported. The extraordinary structural design endows the printed ITO transistors with high device uniformity, outstanding electrical performance, and superior process efficiency. Furthermore, a 10 × 12 coffee-ring structure optoelectronic synaptic transistor array with a high spatial resolution (142 dpi) exhibits durable visual detection and memory behaviors. Meanwhile, an artificial neural network based on the coffee-ring structure synaptic transistors exhibits high accuracy in the recognition rate. These results promise the efficiency of the exceptional printing strategy in developing advanced optoelectronic synaptic transistors for artificial visual systems.