Objective To explore the correlation between triglyceride-glucose（ TyG） index and cognitive impairment in patients with Parkinson's disease（ PD）. Methods The clinical data of 268 PD patients in the Parkinson's Group of the Department of Neurology， Xuanwu Hospital of Capital Medical University from August 2019 to December 2022 were selected for retrospective analysis. Montreal Cognitive Assessment（ MoCA） was applied to assess the subjects' cognitive function. Patients with MoCA ＜ 26（ plus 1 point for illiteracy） were categorized into cognitive impairment group（ 169 cases）， and those with MoCA ≥ 26 were categorized into non-cognitive impairment group（ 99 cases）. General information and laboratory parameters such as triglycerides（ TG）， glycosylated hemoglobin（ HBA1c）， and fasting plasma glucose（ FPG） were collected from the subjects and the TyG index was calculated. The motor symptoms of PD patients were evaluated by the Unified Parkinson's Disease Rating Scale Part Ⅲ（ UPDRS Ⅲ） of the Movement Disorders Society， and their severity was evaluated by using the modified Hoehn-Yahr stage. Quality of life of PD patients was evaluated using the 39-item Parkinson's Disease Questionnaire（ PDQ-39）. Multifactorial Logistic regression was used to analyze the risk factors of cognitive impairment in PD and to construct a Logistic regression prediction model for cognitive impairment. Area under the receiver operating characteristic（ ROC） curve was plotted and the optimal model best cut-off values and sensitivities and specificities were calculated. Results Univariate analysis showed that the differences in age， education， UPDRS Ⅲ score， PDQ-39 score， HBA1c， TyG index， urinary frequency， and diabetes mellitus were statistically significant between the two groups（ all P＜0.05）. Multifactorial Logistic regression analysis showed that age［ OR=1.06， 95%CI（ 1.02， 1.09）］， TyG index［ OR=5.67， 95%CI（ 2.80， 11.46）］， and diabetes mellitus［ OR=2.99， 95%CI（ 1.07， 8.36）］ were independent influencing factors of cognitive impairment in patients with PD. The results of the ROC curve showed that the best cutoff value of the prediction model constructed from the three indicators of age， TyG index， and diabetes mellitus was 0.635， with a sensitivity of 0.692， a specificity of 0.768， and an area under the ROC curve of 0.777. Conclusions Age， TyG index， and diabetes mellitus are independent influencing factors of cognitive impairment in PD. A prediction model consisting of age， TyG index and diabetes mellitus has a good predictive performance for cognitive impairment in patients with PD.