An accurate deterministic of uncertainty parameter is one of the most important aspects of structural uncertainty analysis. However, in actual engineering, it is often difficult to estimate the probability of distribution of uncertainty parameters because of the limited measured data. As such, the information diffusion theory is introduced, and structure parameter uncertainty quantification and propagation method using limited data is proposed. First, the information diffuse theory is used to estimate the probability density function of the limited measured data. Then the random sample is generated using the accept-rejection method and the mean and variance of the uncertainty parameters are obtained by performing the statistical analysis. Second, on the basis of parameter uncertainty quantification, the structural response variability caused by parameter uncertainty is quickly calculated using response surface model. Finally, the proposed method has been validated by a numerical mass-spring system and a set of experimental steel plates. The results show that the variation of the structural response caused by the variation of the parameters can be quantified using the proposed method, without an assumption on the probability of uncertain parameter distribution under the number of measured samples is greater than 20.