Abstract: This paper presents a study on frequency analysis method of hydrological samples with zero values, focusing on the derivation of parameter estimation formulas of Gamma distribution for the nonzero values and full series using maximum likelihood method (ML), the principle of maximum entropy (POME), and probability weight moment method (PWM). We also discuss the relationship between nonzero values and full series and the calculation of design values of hydrological samples. Monte-Carlo simulations are conducted to examine the influence of different sample lengths on the design values calculated using different methods of parameter estimation. Applicability of these methods are evaluated through a case study of six monthly precipitation samples observed in Shaanxi, including the effect and calculation error of fitting large sample events. The results indicate that both the principle of maximum entropy and maximum likelihood method have a good fitting effect, forming a theoretical basis for frequency analysis of the hydrological samples with zero values.