Background Serum uric acid (SUA) has been shown to play an important role in the pathophysiology of mood disorders including 3- and 6-month post-stroke depression (PSD). This study aimed to investigate whether SUA levels on admission were associated with PSD at discharge. Methods A total of 498 stroke patients were consecutively recruited from Tongji Hospital. Clinical and laboratory test data were collected on admission. They were categorized into equal tertiles according to the distribution of SUA and the number of patients. PSD status was evaluated by DSM-V criteria and 17-item Hamilton Rating Scale for Depression at discharge. Results The optimal cut-off points of SUA were: (T1) 80.00 similar to 300.80 mu mo1/L, (T2) 300.81 similar to 391.67 mu mo1/L, (T3) 391.68 similar to 710.0 mu mo1/L. A total of 232 patients (46.59%) were diagnosed as PSD at discharge. Significant differences were found between the PSD and non-PSD groups in SUA tertiles of patients (P = 0.00). After adjustment for conventional confounding factors, the odds ratios of PSD were 5.777 (95% CI = 3.463 similar to 9.637, P = 0.00) for the lowest tertile and 4.153 (95% CI = 2.492 similar to 6.921, P = 0.00) for the middle tertile of SUA, as compared with the highest tertile. In restricted cubic spline regression, continuous SUA showed linear relation with PSD risk at discharge after 300 mu mol/L. Conclusions Lower SUA levels on admission were found to be associated with PSD at discharge and the threshold effect was also revealed. For stroke patients, doctors should pay attention to the baseline SUA for screening high-risk PSD at discharge in clinical practice.