Stock assessment and fisheries management need supporting data that can be collected through fishery-dependent or fishery-independent surveys. The main objective of cost-effective fishery-independent survey design is to collect high-quality data with limited survey cost, and optimization of survey design is often conducted to improve the sampling efficiency. In order to compare the performances of different sampling designs in a fishery-independent survey in estimating abundance indices of multiple target species, four fish species with different spatial distributions, including Larimichthy polyactis, Chaemrichthys stigmatias, Enedrias fangi and Conger myriaster were selected as target fish species. Relative abundance data collected from the bottom trawl surveys conducted in the offshore waters of Shandong Province in 2016-2017 were used to simulate the spatial distributions of relative abundance of target species using Kriging interpolation method. It is assumed that the interpolated relative abundance data were the ‘true’ distribution of four target fish species. The simple random sampling (SRS), the regular systematic sampling (SYS_r), the hexagonal systematic sampling (SYS_h), the stratified random sampling with strata defined by depth (StRS_depth), region (StRS_region), and depth and region (StRS_total) were chosen as the potential sampling designs for estimating abundance index of each fish species at different sample sizes from 20 to 200. The relative estimation error (REE) and relative bias (RB) were used to measure performances of different sampling designs. The accuracy change rate (ACR) less than or equal to 10% was set as the standard for determining the optimal sample size. Computer simulation study was used to compare the accuracy and the precision of different sampling schemes in estimating the abundance indexes. The results showed that the estimation accuracy of three sampling methods was different: simple random sampling < stratified random sampling < system sampling. Except for the systematic sampling, estimates of abundance indices of target fish species in other sampling methods were unbiased. The REE values of target species abundance index estimation in systematic sampling fluctuated irregularly with sample size increasing. The ACR was not suitable to determine the optimal sample size for systematic sampling design. The REE for the stratified random sampling was slightly higher than that of systematic sampling, but it showed higher stability. The stratified random sampling was the best sampling method. The stratified random sampling StRS_total could reduce the REE value compared with other two stratified sampling schemes. It was the best stratification scheme. The optimal number of stations for design StRS_total in four seasons could be set at 80. This study also offered references to determine the optimal sampling design for multispecies fishery-independent surveys in other waters.