Abstract: Fish growth exhibits non-stationary characteristics in response to multiple factors, including climate change, environmental stress, and fishing pressure. Traditional stock assessments are generally based on stationarity assumptions, neglecting temporal growth variability, which may lead to biases in the estimation of key management reference points and consequently affect fisheries management decisions. In this study, the Eastern Atlantic skipjack tuna (Katsuwonus pelamis) was selected as a case study. Sea surface temperature was incorporated into the dynamic modeling of growth parameters to quantify temporal growth variability, and the impacts of such variability on stock assessments were further investigated. The results show that models incorporating temporal growth variability estimated overall lower and more stable spawning stock biomass. In contrast, neglecting temporal growth variability tended to overestimate spawning stock biomass and underestimate fishing mortality, resulting in a negative bias (−38.24%) in the maximum sustainable yield (MSY). Meanwhile, the relative spawning stock biomass (SSB_cur/SSB_MSY) and relative fishing mortality (F_cur/F_MSY) showed significant negative and positive biases, respectively. This study demonstrates that incorporating temporal growth variability into fisheries stock assessments helps to more comprehensively reflect the potential population dynamics and provides more robust scientific support for fisheries management, thereby avoiding the risk of overexploitation caused by misjudgment.