The study examined 419 patients with newly diagnosed pulmonary nodules at least 19 months after confirmed COVID-19 infection (January 2020–December 2024). Among them, 210 patients (50.1%) had persistent nodules at 6 months. Independent risk factors for persistence were ≥ 4 hospitalizations, previous tuberculosis, larger maximum nodule diameter (OR increase per mm: 1.121, 95% CI: 1.074–1.170), and vascular convergence sign on CT. A predictive model integrating radiomic, clinical, and epidemiological variables selected four key predictors using the LASSO method and achieved an AUC of 0.728 (bootstrap corrected AUC of 0.717). Decision curve analysis demonstrated clinical benefit at threshold probabilities of 50–83%. The model can help optimize the follow-up of patients at higher risk.