The effect of long‐term continuous positive airway pressure (CPAP) treatment on apnea–hypopnea index (AHI) after CPAP withdrawal remains unclear, especially in obstructive sleep apnea (OSA) patients screened from the population. To examine that, 1241 civil servants who participated in the annual physical examination were screened for OSA between September and December 2017. Screened OSA firstly underwent 1‐week CPAP adherence assessment. Then, patients with good CPAP adherence would be freely provided CPAP to continued treatment. All OSA patients were followed for 2 years. At study end, all OSA patients underwent home sleep testing (HST) again within 1 week of CPAP withdrawal. The effect of 2‐year CPAP treatment on OSA severity was investigated by using linear regression and multinominal logistic regression. In total, 103 OSA patients were screened, including 41 cases (39.8%) in CPAP treatment group and 62 cases (60.2%) in non‐CPAP treatment group. At 2‐year follow‐up, compared with baseline, in CPAP treatment group, following CPAP withdrawal, a significant decrease in AHI was observed in patients with severe OSA (P = 0.014); in non‐CPAP treatment group, a significant increase in AHI was observed in patients with moderate OSA (P = 0.028). After adjustment for confounding factors, multivariate linear regression showed that △AHI was negatively associated with CPAP treatment (β = −4.930, 95% confidence interval [CI] [−9.361, −0.500], P = 0.030). Multinominal logistic regression showed that the AHI of patients not treated with CPAP tended to be unchanged or worsened with the AHI improvement group as a reference (OR [odds ration] [95% CI], 4.555 [1.307, 15.875], P = 0.017; 6.536 [1.171, 36.478], P = 0.032). In conclusion, active OSA screening and long‐term CPAP intervention may improve the severity of severe OSA patients following short‐term CPAP withdrawal in the general population.