Abstract: Fisheries stock assessment paradigms include detailed process-oriented Russell’s model and simpler biomass dynamic models, and the latter integrated individual population processes. This study evaluates the performance of a Beverton–Holt-Driven production function (BHDPF) model in assessing the Chinese anchovy stock (Engraulis japonicus). This model more explicitly adheres to Russell’s four-process framework and is different with the flexible traditional biomass dynamic model (the Pella–Tomlinson form). Using an age-structured operating model based on Russell’s model and informed by anchovy life-history data, we simulated population dynamics under both Beverton–Holt and Ricker stock-recruitment relationships (SRRs) across four fishing mortality (F) scenarios. The BHDPF and Pella–Tomlinson models were then modified to fit the simulated catch and catch per unit effort (CPUE) data. Their performances in estimating biomass, depletion, and F were compared. The results demonstrate that the BHDPF model consistently outperformed the Pella–Tomlinson model, yielding more accurate and precise estimates with lower bias and variability across most scenarios and SRR types. Both models performed worse under Ricker SRR. This study underscores the importance of integrating more biological realism, such as explicit stock-recruitment dynamics, into assessment models for data-limited fisheries. The BHDPF thus represents a viable alternative for practical application.