The size spectra of fish communities can reflect the characteristics of their biomass, abundance, and size structure, and these parameters can be used to explore the effects of environmental changes and human disturbance on the ecosystem. Therefore, it is important to construct reasonable size spectra of fish communities to study their community structures. However, there are differences in the quantity and quality of data collected from different survey designs, which further affects the construction of the size spectra of fish communities. This simulation study examined the effects of sampling design on the estimation of parameters of normalized biomass size spectra of fish communities using a computer resampling method based on data collected from seasonal bottom trawl surveys in the southern waters off the Shandong Peninsula from 2016 to 2017. Simple random sampling (SRS), stratified random sampling with strata defined by depth (StRS_depth), and stratified random sampling with strata defined by size (StRS_size) were chosen as potential sampling designs for sample sizes ranging from 5 to 60. The relative estimation error (REE) and relative deviation (RB) were used to measure the performance of the different sampling designs. The results showed that under the three sampling methods, REE values of the slopes of the normalized biomass size spectra of the fish communities decreased with an increase in the sample size. As the sample size increased, the range of absolute values of RB decreased and gradually approached zero. Under the three sampling methods, the distribution of simulated values of the slopes of the normalized biomass size spectra of fish communities became more concentrated, and the estimation accuracy improved with an increase in sample size. StRS_size performed better compared with those of the other two sampling methods, which had lower REE and absolute RB values. In addition, different seasonal data showed different effects on the slope estimation of the normalized biomass size spectra of fish communities. Considering the sampling effects of different sample sizes during the four seasons under the three sampling methods, StRS_size was the optimal sampling method suitable for the estimation of biomass size spectra parameters of fish communities, and the number of sampling stations could be optimized.