Purpose To explore the relationship between transcranial sonography (TCS) echo changes in the midbrain substantia nigra (SN), lentiform nucleus (LN), and midbrain raphe (BR) and early diagnosis of Parkinson"s disease (PD). "s disease (PD). METHODS Retrospective analysis of 231 patients seen at the Army Specialty Medical Center from 2020 to 2023, all of whom underwent TCS, divided the 231 patients into PD (+) and PD (-) groups, analyzed the echogenic characteristics of the substantia nigra, the nucleus pulposus, and the nucleus of the median suture, and performed assignment modeling. Results Comparison of gender and clinical symptoms (tremor) between the two groups showed no statistically significant difference (P > 0.05); comparison of age, substantia nigra, nucleus pulposus, and nucleus of the middle suture showed statistically significant differences (P < 0.05). One-way regression analysis revealed that age, substantia nigra, nucleus accumbens, and nucleus pulposus were independent correlates of the diagnosis of PD, whereas gender and clinical symptoms (tremor) were not independent correlates of the diagnosis of PD. A logistic regression prediction model based on the above four variables was constructed: Y=-7.338+0.038×Age+0.991×SN+1.765×BR+1.076×LN. and the ROC curve was plotted, with an AUC value of 0.851, which attained a moderately high predictive value, and a sensitivity of 79.6%, specificity of 80.5%, and an approximate Den index of 0.601, and a cutoff value of 0.393. DCA curves were plotted based on this model, and in the DCA curves, there was room for a large clinical benefit for patients at almost all threshold probabilities, suggesting that there can be a net clinical benefit for patients when this model is used for PD prediction. Conclusion Age, substantia nigra, nucleus pulposus, and nucleus intermedius are independent correlates of the diagnosis of PD, and the prediction model constructed on the basis of these factors has a moderately high predictive value, and the patients have a good net clinical benefit when used for the prediction of PD.