Background: The Montreal Cognitive Assessment Basic scale (MoCA-B) is more sensitive than the Mini-Mental State Examination (MMSE) for detecting mild cognitive impairment due to Alzheimer’s disease (AD). To explore the diagnostic efficacy of the Chinese version of the MoCA-B against the MMSE for post-stroke cognitive impairment (PSCI). Methods: Eighty four patients with acute cerebral infarction were grouped into a post-stroke cognitive normal (PSCN) or a PSCI group based on their scores on the Clinical Dementia Rating scale (CDR), the gold standard for diagnosing PSCI. They were evaluated by using the MMSE and MoCA-B scales, then the area under the receiver operating characteristic (ROC) curve (AUC) was used for evaluation. Results: Most factors of the MoCA-B were significantly different between the two groups, and the PSCN group completed the MoCA-B faster (p < 0.05). The AUC analysis showed that for the MoCA-B with a cut-off total score of 23, sensitivity = 85.71%, specificity = 61.22%, Youden’s J Index = 0.469, and AUC = 0.832. For the MMSE with a cut-off total score of 25, sensitivity = 70.59%, specificity = 93.75%, Youden’s J Index = 0.643, and AUC = 0.885. The AUC of the MMSE was higher than that of the MoCA-B (p > 0.05). The MoCA-B had greater sensitivity and negative predictive value than the MMSE. When considering the cutoffs for identifying mild cognitive impairment (MCI) across different education levels, the MoCA-B had a higher positive rate for PSCI identification (51.2% vs 25%, p < 0.001), indicating that the MoCA-B is suitable for identifying PSCI. Conclusion: The MoCA-B demonstrates higher sensitivity and negative predictive value compared with the MMSE in the screening of post-stroke cognitive impairment patients.