A model was constructed to estimate Primary production (PP) and examine the effect of the dominant phytoplankton group on PP, using a dataset collected in 2019 in the South China Sea (SCS) based on phytoplankton absorption coefficient at 443nm [a(443)] and photosynthetically active radiation (PAR). There was a significant log-log linear correlation between PP and the product of a(443) and PAR (a(443)×PAR), with an adjusted R of 0.64. The model was validated using K-fold cross-validation and an in situ dataset collected in 2018 in the SCS basin. The results showed that the model had good generalisability and was suitable across marine environments, including basin, coastal, and offshore areas. The model was more sensitive to changes in PAR than changes in a(443). Phytoplankton in the diatom-dominant and haptophyte-dominant clusters were in the light-limited stage, and their PP values increased with increasing a(443)×PAR. However, Prochlorococcus-dominant samples exhibited photoinhibition, and the PP values decreased with increasing a(443)×PAR, likely due to their bio-optical characteristics. The model’s predictive power was related to the photo-physiological state of dominant phytoplankton, which performs well in light-limited conditions but not in cases of massive photoinhibition. This study provides insight into the development of phytoplankton-specific a-based PP models.