This paper presents a framework for quantifying uncertainty in per-recruit analysis for small yellow croaker (Larimichthys polyactis) fisheries in China, in which credible estimates of life history parameters from Bayesian inference were used to generate the distribution for a quantity of interest. Small yellow croakers were divided into five spatial groups. The status of each group was examined using a yield-per-recruit (YPR) model and a spawning stock biomass-per-recruit (SSBPR) model. The optimal length at first capture (Lcopt) was proposed to recover the biomass. The maximum observed age in the current stocks (3 years) and the maximum recorded age (≥20 years) were adopted in per-recruit analysis. Our results suggest that the framework can quantify uncertainty well in the output of per-recruit analysis for small yellow croaker. It is suited to other fish species. The SSBPR at FMSY (SSBPRMSY) is a better benchmark than the spawning potential ratio (SPR) at FMSY because SSBPRMSY had a unimodal distribution. The SSBPR analysis can lead to a more conservative Lcopt than the YPR analysis. The key factor influencing the assessment conclusions may be the growth parameters rather than the natural mortality rate for a stock with a younger maximum age. Overfishing likely occurred for all groups and recruitment overfishing may not occur if the maximum age is maintained at 3 years. Increasing lengths at first capture to the recommended values can help this population recover. However, Fcur is too high for small yellow croakers to attain the maximum recorded age. Both reducing fishing mortality rate and increasing length at first capture are needed to attain the maximum recorded age.