Fishery-independent survey is commonly used for collecting high quality biological and ecological data and provides vital parameters for fish stock assessment and management. For fishery-independent surveys, low cost and high precision as a result of reduced sample size may lead to a much greater reduction of survey mortality for the already depleted fishery populations and the ecosystem of fragile estuarine habitats, so that optimization of sampling design is essential for acquiring cost-effective sampling efforts. Computer simulation is usually applied to evaluating sampling designs in identifying an optimal sampling effort to achieve the goals of a survey program. In this study, the effect of sample size on estimation of abundance index was examined for fishery species including Cynoglossus joyneri, Pholis fangi, Chaeturichthys stigmatias, Loligo sp., Oratosquilla oratoria, Charybdis japonica and small-sized fish species in Clupeidae and Engraulidae in a cluster sampling survey based on the data collected from bottom trawl surveys conducted in August, October 2013 and February, May 2014 in the Yellow River estuary and its adjacent waters. The relative estimation error (REE), relative bias (RB), coefficient of variation (CV), and accuracy change rate (ACR) were used to measure the performances (accuracy and precision) of different sampling efforts. In general, these indices including REE, CV and ACR showed similar trends, decreasing with sample size initially and then becoming stable after certain sample sizes during all the four seasons. Most of RB did not show consistently increasing or increasing trends with sampling size. The absolute value of RB of C. japonica, O. oratoria and P. fangi etc showed an obvious increase when the sample size reduced from 3 to 1. The optimal sampling efforts were different for the selected species because the target species had different spatial variabilities. The effects of reducing sampling effort on the precision of abundance index estimation were low when the species had low spatial variabilities. Therefore, the optimal sampling effort for multiple species fishery resources surveys should consider the spatial and temporal distribution patterns of the target species.