The objective of this observational study is to evaluate the accuracy and feasibility of an artificial intelligence-based multimodal cognitive screening system in early identification of diabetes-related mild cognitive impairment (MCI) among type 2 diabetes mellitus (T2DM) patients through a 3-5 year follow-up. It also aims to analyze the correlation between diabetic metabolic indicators (such as glycemic variability and HbA1c levels) and cognitive function changes, thereby determining the value of this intelligent screening system in early detection and intervention of cognitive impairment in T2DM patients, which constitutes the key focus of this research.