Accurate understanding of changing population dynamics associated with climate change is critical for managing highly migratory fish species. However, long-term data on population dynamics and the resulting recruitment variability is still lacking for many species, making it difficult to predict and integrate the effects of ocean warming into management schemes. In this study, high-resolution stable oxygen isotope (δ18O) analysis was performed on the otoliths of adult Pacific bluefin tuna Thunnus orientalis using secondary ion mass spectrometry (SIMS) to determine the natal origin of an individual fish. The core δ18Ootolith corresponding to the larval stage greatly varied among the individuals, indicating that the larvae experienced a wide range of thermal environments. The non-hierarchical cluster analysis performed on the core δ18Ootolith grouped fish into those with higher δ18Ootolith (lower temperature) and those with lower δ18Ootolith (higher temperature), most likely representing relative temperature difference experienced between fish born in the Sea of Japan and in the Nansei Islands area. The Nansei Islands area cluster showed more variability in the early otolith growth indicating a longer spawning season, which is consistent with the observed longer spawning duration in this area. The absolute temperature estimates based on the SIMS-measured core δ18Ootolith were significantly higher than those expected from sea surface temperature data, suggesting the effects of matrix-related bias on the temperature offsets. The relative temperature difference, however, matched well with the known spawning temperature range of the two spawning grounds. The recruitment contribution from each spawning ground (all year-classes pooled, n = 51) was 45% in the Sea of Japan and 55% in the Nansei Islands area. Overall, this study demonstrated the effectiveness of SIMS δ18Ootolith analysis for investigating the natal origin of fish and its potential application in fish population dynamics studies.