The latest consensus has changed CYP2D6 genotyping among Chinese population, while its impact on metoprolol tolerance and adverse events in elderly Chinese patients with cardiovascular diseases remains unclear. In this study, we prospectively included elderly patients who started metoprolol treatment for cardiovascular indications. According to the latest consensus on CYP2D6 genotype-to-phenotype translation, the patients were categorized as normal, intermediate, or poor metabolizers (NMs, IMs, or PMs, respectively) by detecting the presence of the CYP2D6*1, *2, *5, *10, and *14. Logistic regression model was used to analyze the correlation between the CYP2D6 phenotype and incidence of adverse events, which were assessed over a 12-week period. In this study, there were 651 (62.7%) NMs, 385 (37.1%) IMs, and 3 (0.3%) PMs. After 12 weeks of follow-up, compared with NMs, IMs had the lower maintenance dose [50.0 (25.0–50.0) mg/day vs. 25.0 (25.0–50.0) mg/day, p < 0.001] and lower weight-adjusted maintenance doses (0.52 ± 0.25 mg/day/kg vs. 0.42 ± 0.22 mg/day/kg, p < 0.001), and had higher incidence of postural hypotension (6.0% vs. 10.9%, p = 0.006), bradycardia (21.5% vs. 28.6%, p = 0.011), asystole (0.8% vs. 3.1%, p = 0.009) and syncope (2.0% vs. 6.2%, p = 0.001). In logistic regression model, the overall incidence of adverse events was 1.37-fold larger in IMs than in NMs (odds ratio = 1.37, 95% confidence interval = 1.05–1.79, p = 0.021). We conclude that IMs have lower tolerance and higher incidence of metoprolol-related adverse events than NMs in elderly Chinese patients with cardiovascular diseases. CYP2D6 genotyping is justifiable in elderly patients to minimize the risk of adverse events and ensure the benefits of metoprolol.