We examined the associations of polygenic risk score (PRS) with Alzheimer's disease (AD) and plasma biomarkers in the Chinese population. This population-based study used baseline data from MIND-China (2018; n=4873) and follow-up data from dementia-free individuals (2014-2018; n=2117). We measured AD-related plasma biomarkers in a subsample (n=1256). Data were analyzed using logistic and Cox regression models. We developed PRS with (PRSAPOE) and without (PRSnon- APOE) apolipoprotein E (APOE) gene. In the longitudinal analysis, PRSAPOE was associated with a multivariable-adjusted hazards ratio of 1.91 (95% CI=1.13-3.23) for AD. PRSAPOE in combination with demographics yielded discriminative (area under the curve [AUC]) and predictive(C-statistic) accuracy of 0.80 (95% confidence interval [CI]=0.77-0.84) and 0.80 (0.77-0.82), respectively. PRSnon- APOE showed an association with AD risk similar to PRSAPOE. PRSAPOE, but not PRSnon- APOE, was associated with reduced plasma Aβ42/Aβ40 ratio and increased Neurofilament light chain (NfL) (p<0.05). The PRS with and without APOE gene, in combination with demographics, shows good discriminative and predictive ability for AD. The AD-related pathologies underlie AD risk associated with PRSAPOE. The PRSAPOE and PRSnon- APOE were associated with AD risk in the Chinese population. The PRSAPOE and PRSnon- APOE, in combination with demographics, showed good discriminative and predictive ability for AD. The AD-related pathologies underlie the AD risk associated with PRSAPOE but not PRSnon- APOE.