Optical coherence tomography angiography (OCTA), as a non-invasive three-dimensional fundus vascular imaging technique, has significant advantages in the diagnosis and follow-up of eye diseases such as diabetic retinopathy and age-related macular degeneration. However, the existence of OCTA image artifacts has seriously affected its clinical application. These artifacts are caused by various factors such as image acquisition, internal characteristics of the eyeball, eye movement and image processing, such as weak signals, blinking, defocusing, bands, tilting, occlusion, exposure, projection, movement and layering, leading to vascular quantization deviation, lesion blurring and image distortion, thereby reducing the accuracy of clinical diagnosis. To address this issue, researchers have proposed a variety of correction strategies, including enhancing signal strength, optimizing equipment, developing algorithms to identify and eliminate shadow artifacts, using hardware or software methods for motion correction, and employing deep learning algorithms for image quality assessment and artifact removal. Constructing a unified and systematic framework for artifact cognition and processing is crucial for enhancing the reliability of OCTA diagnostic results and will drive the level of ophthalmic diagnosis and treatment to a new height.