Chinese shrimp (Fenneropenaeus chinensis) in stocking program was used an example, to illustrate the uncertainty associated with empirically estimated natural mortality on the dynamics and structure of fish populations. Two empirical formulas based on the growth parameters were used to estimate the natural mortality rates which were then compared to the natural mortality estimated based on catch data. The differences in three natural mortalities over time and their impacts on sex ratio were evaluated. Our analyses showed that estimation methods could greatly affect the estimation of natural mortality. Although the mating mortality was included in the natural mortality obtained based on the catch data (Ye et al., 1987), the M of young fish has been underestimated at early releasing stages. The M was overestimated with the empirical formula proposed by Chen & Watanabe (1989), and the sex ratio was 4.44:1 at the end of one-year old China shrimp. The M was underestimated at the early releasing and overestimated at the steadily growth stage with the Gislason's method, and the sex ratio was 2.22:1 at the same time. At the beginning and end of fishing, the amounts of resources and catches based on Ye et al method were over 2.42 and 2.87 times the amounts estimated using the Gislason's empirical formula, and about 76.25 and 102.50 times of the values estimated with Chen and Watanabe's method. These data suggest that the empirical methods selected to estimate the natural mortality need to be based on the most rigorous criterion, and when calculating the natural mortality gender-specifically by empirical methods, sex parameter should be included to ensure that the results are biologically meaningful.