Continuous glucose monitoring is transforming analysis by using sweat, a less invasive and more accessible fluid, to provide more accurate readings during exercise. However, obtaining accurate results remains challenging due to its acidity, which can disrupt the alkaline conditions required for non-enzymatic sensing. Herein, we developed a high-surface-area carbon nanofiber (CNFs) composite functionalized with polypyrrole (PPy) and combined with a graphene oxide (GO) composite for non-enzymatic electrochemical glucose sensing in sweat. The composite was fabricated via a sequential drop-casting process using a metal-azolate framework (MAF) for its high electrochemical stability, thereby enabling OH-compensation and ensuring a continuous supply of hydroxyl (OH) groups from GO, compensating for any loss of OH caused by sweat acidity. The CNFs/PPy/GO/MAF composite works efficiently across pH levels 13, 12.5, and 12, with detection limits of 0.033, 0.043, and 0.0145 µM (S/N = 3), respectively. The composite shows high sensitivities of 663, 259.6, and 160.85 μA/(mM·cm²) within the linear ranges of 0.003 to 1.6, 0.003 to 1.4, and 0.003 to 0.7 mM at pH = 13, 12.5, and 12, respectively. Additionally, the composite exhibits high selectivity, stability under decreasing pH, and good repeatability and reproducibility. It yielded sweat analysis results from real samples that closely matched blood glucose levels measured by a glucometer at all tested pH values of 13, 12.5, and 12. This confirms that the composite offers accurate, reliable, real-time, non-invasive glucose detection under typical alkaline conditions.