The role of climate variability in determining the fluctuations of fish populations had been a traditional problem in ecology. In this paper, we studied the role of the Southern Oscillation Index (SO) and the Pacific Decadal Oscillation (PDO) on the population dynamics of the western stock of the skipjack tuna Katsuwonus pelamis. Our analysis was based in three sequential steeps: a diagnostic approach to deduce what kind of population dynamic model should be more appropriate, the modelling of capture per unit of effort data through a logistic model, and the use of population dynamic theory for analyzing the effect of exogenous perturbations. We find that direct and one-year lagged negative PDO effects and one-year lagged negative SO effects were needed to explain annual tuna fluctuations. Models including the combined effects of these climatic indexes explain 80% of the variance in tuna fluctuations. In addition, these models provided very accurate predictions of independent skipjack tuna observed dynamics. This result is encouraging because the inherent variability in CPUE data and the not well determined link between climate and ecological processes. Finally, this study demonstrates that simple models can offer reasonable explanations and accurate predictions of tuna fluctuations, provided they are based on a sound theoretical framework.