This 24-week randomized controlled trial compared a continuous glucose monitoring (CGM)-guided personalized postprandial-targeting (PPT) diet with the standardized Chinese Diabetic Diet (CDD) in 175 Chinese adults with prediabetes or unmedicated type 2 diabetes (T2DM). The PPT diet dynamically adjusted macronutrients composition based on individualized glucose responses, whereas the CDD followed national guidelines. Compared with the CDD group, the PPT group achieved significantly greater reductions in HbA1c [-0.46% vs. 0.06%; between-group difference: -0.4% (-0.55, -0.25)%, P < 0.001], fasting plasma glucose [-1.12 vs. -0.74 mmol/L; -0.37 (-0.64, -0.11) mmol/L, p = 0.007], and 2-h oral glucose tolerance test [-3.52 vs. -2.47 mmol/L; -1.04 (-2.13, -0.04) mmol/L, p = 0.028]. PPT also reduced postprandial glucose exposure [incremental area under the curve: -101.24 vs. -59.39 mmol/L·min; between-group difference: -42.86 (-70.4, -16.08) mmol/L·min, P < 0.003], accompanied by a greater reduction in carbohydrate consumption (244.0→148 g/day vs. 240→181.4 g/day for CDD), improved lipid profiles (LDL-C: -1.02 vs. -0.23 mmol/L; triglycerides: -0.67 vs. -0.1 mmol/L; HDL-C: + 0.31 vs. + 0.13 mmol/L; all P < 0.001), and more pronounced weight loss [-6.83 vs. -0.55 kg; -5.97 (-7.2, -4.74) kg, P < 0.001]. Liver enzymes improved significantly in the PPT group (ALT: -6.23 vs. -0.97 U/L, p = 0.004; GGT: -8.41 vs. -1.04 U/L, P < 0.001), as did self-management scores. These findings demonstrate that CGM-guided personalized nutrition can substantially improve glycemic control, lipid metabolism, and overall metabolic health, providing a promising and potentially scalable strategy for high-risk populations. However, the long-term sustainability and cost-effectiveness of this approach warrant further study.