Objective To explore the predictive value of combining type 2 innate lymphoid cell （ILC2） proportion and soluble suppression of tumorigenicity 2 （sST2） protein level for postoperative recurrence in refractory chronic rhinosinusitis with nasal polyps （CRSwNP）. Methods We retrospectively reviewed the clinical data of 324 patients with refractory CRSwNP admitted to the hospital from January 2021 to December 2023， who constituted the training set. All patients underwent endoscopic sinus surgery and were followed up for 12 months， with examinations every 3 months. Based on recurrence status， they were divided into a recurrence group （n = 125） and a non-recurrence group （n = 199）. Clinical data， ILC2 proportion， and sST2 levels were compared between the two groups. Univariate and multivariate logistic regression analyses were performed to identify factors influencing postoperative recurrence. The predictive value was assessed using receiver operating characteristic （ROC） curves. Additionally， 138 patients meeting the same criteria admitted from January to September 2024 were enrolled as the validation set. A nomogram prediction model was constructed and evaluated using ROC curves， calibration curves， and decision curve analysis in both sets. Results The recurrence group showed significantly higher age， olfactory function score， Lund-Mackay score， Davos score， neutrophil count （NEUT）， ILC2 proportion， and sST2 level compared to the non-recurrence group （all P < 0.05）. Significant differences were also observed in allergic rhinitis history， rhinosinusitis subtype， and operation duration （all P < 0.05）. Multivariate logistic regression identified type Ⅲ rhinosinusitis， ILC2 proportion and sST2 level as independent early postoperative predictive indicators for recurrence （both P < 0.05）. The combined prediction of ILC2 and sST2 demonstrated superior value for postoperative recurrence compared to either marker alone， with a higher area under the curve （AUC） （Z_{ILC2~combined} = 3.367， Z_{sST2~combined} = 4.464， P < 0.05）. The nomogram model showed good predictive performance， with AUCs of 0.88 （0.84 ~ 0.93） and 0.78 （0.68 - 0.89）， The calibration curve indicated good agreement between predicted and actual outcomes （P > 0.05）. Decision curve analysis revealed a high net benefit at threshold probabilities between 40% and 90%. Conclusion The combination of ILC2 cell proportion and sST2 protein level holds significant predictive value for postoperative recurrence in refractory CRSwNP， providing a potential reference for clinical risk stratification and targeted intervention.