The current paper presents a perturbation-based stochastic eigenvalue buckling analysis of thin plates using element free Galerkin method. Spatial variation in Young's modulus is modeled as a homogeneous random field and moving least square-based shape function method is employed for discretizing the random field. Perturbation method is used to evaluate the statistics of buckling loads. Numerical examples wherein rectangular plates with different boundary conditions are solved and the statistics obtained are compared with those calculated using Monte Carlo simulation. Different parametric studies are also conducted. The results obtained from perturbation method are found to be reasonably accurate for coefficient of variation (CV) values less than 20% for random fields with normal distribution. Further it is observed that for random fields with lognormal distribution, the proposed method produces reasonably accurate results up to a CV of 30%. © 2021 World Scientific Publishing Company.