Many ecosystem models have been developed recently for simulating oceanic biogeochemistry, but most of the previous models assumed a constant diffuse attenuation coefficient and were not spectrally resolved. We developed a bio-optical model that explicitly represents two phytoplankton (picoplankton and diatoms) and two zooplankton (micro- and meso-zooplankton) functional groups, as well as multiple nutrients and non-algal particles (NAP). The hourly surface ocean color field and subsurface light field is calculated by a radiative transfer model (Ecolight) using relationships linking the ecosystem model constituents with inherent optical properties. Phytoplankton absorption is partitioned to the different functional groups. Spectral particulate backscattering is a function of particulate organic carbon (POC) concentration, but has different relationships among phytoplankton species. We apply the model to the equatorial Pacific upwelling region, and the model is capable of reproducing measured optical properties and key biogeochemical processes in this region. The model results show large contribution of NAP to total scattering or attenuation (>50% at 660nm) and their small contribution to particulate absorption (<20% at 440nm), and remarkable contribution of picoplankton to total phytoplankton absorption (>95% at 440nm). All these model results are consistent with the field data in the surface waters. Primary production, a crucial variable for assessing oceanic carbon cycling, is estimated directly from modeled carbon fixation and is compared with 14C measurements. This study suggests that optical properties play an important role in identifying and reducing uncertainties in ecosystem models, and provide constraints for determining variables such as NAP and POC concentrations and related parameters that are difficult to measure directly and continuously in time and space.