单一种群是目前渔业资源评估的基本假设,但渔业资源常由多个地方种群或产卵种群组成,并且种群间存在交流,构成复合种群。根据复合种群概念,以东、黄海日本鲐为例,对其12种种群动态情况进行了模拟。利用模拟所得的数据及剩余产量模型,分别分析了在复合种群、两独立种群及单一种群假设下所设置的10种评估管理方案,结果表明:(1)基于复合种群假设的评估管理方案与模拟的种群动态一致,在单位捕捞努力量渔获量(CPUE)观测误差较小情况下,该方案为最佳方案,可获得最大可持续产量,但随CPUE观测误差增大,该方案种群灭绝率增大,管理效果随之退化。(2)基于两独立种群假设的评估管理方案均使资源过度开发,不利于资源可持续利用。(3)在单一种群假设下,选择不同CPUE作为资源指数和采用不同捕捞量分配方法的评估管理方案存在过度捕捞和开发不足两种状况,其管理效果受种群本身参数及空间交换率等因素的影响而不同;若采用的CPUE反映部分种群动态信息,则其评估管理方案至少在一种模拟情况下出现种群100%灭绝;若CPUE能反映整个种群资源量的动态变化,且捕捞量能按种群的空间结构进行分配,则管理效果与(1)类似,但不能获得最大可持续产量,若忽略种群的空间结构影响而均匀分配捕捞量,则至少在一种模拟情况下出现种群100%灭绝。据此,对于复合种群的管理,建议:(A)如果种群数据收集及数据精度能得到保证,该资源的评估与管理应基于复合种群假设;(B)如果目前收集种群数据存在较大困难,且CPUE数据存在较大误差,则可采用单一种群假设,但必须设定更保守的捕捞量和采用基于种群空间结构的总许可渔获量(TAC)管理方案;(C)在制定渔业管理政策时,应结合种群生态、数据、模型假设及参数估计方法等方面的不确定性对管理控制规则进行系统的管理策略评价以避免风险。;A unit stock is a basic assumption in fishery stock assessment. However, the stock always consists of local subpopulations or spawning subpopulations, among which there is some degree of exchange of individuals. Thus, the stock is commonly referred to as meta-population. We simulated two subpopulations of chub mackerel (<em>Scomber japonicus</em>) in the East China Sea and Yellow Sea based on the meta-population concept by setting up 12 scenarios. Using the data derived from the simulations and the surplus production model, we evaluated the effect of stock management based on 10 schemes, which involved different assumptions regarding the spatial structure of the stock and management strategies. The spatial structure of the stock was represented as 1) a meta-population, 2) two independent subpopulations or, 3) a unit stock. For management strategies, we focused on two methods of allocating the catch among the subpopulations: 1) assigning the catch in proportion to the biomass of each subpopulation or 2) assuming that the catch was uniform among the subpopulations by ignoring the spatial structure of the stock. Our results showed that (1) the maximum sustainable yield and the biomass at maximum sustainable yield could be achieved when there was little uncertainty in the CPUE and when the scheme was based on a meta-population assumption. However, the extinction probability of the stock increased with degradation of quality of the CPUE. (2) If we assumed there were two independent subpopulations by ignoring the exchange among the subpopulations, the stock was overfished and was not sustainable. (3) If the assessment and management was based on a unit stock assumption, the stock was either overfished or underexploited, depending on which CPUE was used as the abundance index of the stock, the methods of allocating catch along the subpopulations, and biological parameters such as the carrying capacity, intrinsic growth rate, and migration matrix. If the CPUE time series partially represented the trend of the biomass of the stock, for example by only representing the biomass of one of the subpopulations, there was a 100% probability of extinction under a certain situations. If the CPUE correctly reflected the dynamics of the biomass of the whole stock and the catch was assigned according to the biomass of each subpopulation, the effect of stock management was similar to (1), but it was suboptimal because the maximum sustainable yield could not be achieved. However, if the catch was allocated by ignoring the spatial structure of the stock, there was also a 100% probability of extinction under some situations. Therefore, to perform stock assessment and management for meta-populations, we recommend: (A) if enough information regarding subpopulations for stock assessment can be collected and there is little uncertainty in the observed data, the assessment and management of the stock should be based on a meta-population assumption. Otherwise, (B) a unit stock can be assumed but a more conservative management strategy should be adopted and a TAC (Total Allowed Catch) based on the spatial structure of the stock, i.e., the catch is allocated according to the abundance of the subpopulations, is needed to keep the fishery sustainable. (C) Before making a decision, the assessors or managers should conduct a management strategy evaluation to minimize management risk by evaluating the rules of management and taking into account the uncertainty in our understanding of the ecology of the species, collecting data, and assuming or estimating the parameters of the stock assessment model.