Uncertainties in material properties and geometric dimensions are inherent in the fabrication of thin-walled structures, which significantly influence the load-bearing capacity and reliability. To ensure structural safety, these factors need to be incorporated into the structural stochastic analysis. Specifically, due to the small amplitude of geometric dimensions, precise geometric modeling and analysis are necessary[1]. Leveraging the advantage of isogeometric analysis (IGA) in eliminating geometric discretization errors in modeling[2], this study proposes a perturbation stochastic isogeometric analysis (PSIGA) method based on a nested geometric model to predict the stochastic responses of thin-walled structures under material properties and geometric dimensions. In modeling process, the nested geometric model is adopted to ensure model accuracy and watertightness, securing numerical analysis precision. In uncertainty representation process, a random field modeling method is developed. Uncertainties of structures are represented as random fields via the Karhunen–Loève (K-L) expansion, which are then assembled into the stochastic static and buckling equations and incorporated into the stochastic isogeometric analysis framework[3]. In analysis process, a first-order perturbation method is employed to solve these stochastic equations and derive the statistical characteristics of stiffened panel responses. Moreover, a sensitivity analysis is conducted to quantify the relative influences of individual uncertainties on the structural behavior. The calculation accuracy of the proposed PSIGA method under multiple uncertainties are demonstrated through several examples, with the Monte Carlo simulation (MCS) results serving as the benchmark. The outcomes demonstrate that the proposed PSIGA method efficiently and accurately captures the stochastic behavior of stiffened panels, providing a robust tool for engineering structural safety evaluations.