This study suggests a new approach for ocean color algorithms (OCAs) in Case 2 water using bio-optical field measurements in the Western Arctic Ocean (WAO). Relationship between the inherent optical properties (IOPs) and the apparent optical properties (AOPs) on the sea surface were highly correlated through remote sensing reflectance, Rrs(λ). This result indicates that the radiance and irradiance can statistically reflect the effect of combined the absorption and backscattering of particulates. Based on the AOPs data, estimated surface chlorophyll a concentrations (chla) using SeaWiFS global algorithm, OC4V4 and the tuned for the Arctic, Arctic OC4L, showed much different from in situ chla including 25% and 30% errors, respectively. Absorption coefficients of suspended material and colored dissolved organic matter were also high in addition to high values of Rrs(λ) spectra from 510nm to 565nm, which indicates that the waters is classified into Case 2 water. These results demonstrate that in Case 2 water, current OCAs would be not applicable, and hence a new idea for developing OCA is required. Here as the approach, relationship between in situ chla and Rrs(555)/Rrs(510) is used to identify waters as spectral classified water groups. Especially in Case 2 water, with empirically derived Rrs(555)/Rrs(510) to be 0.75, when tuned intercepts with fixed regression coefficient for all data is applied on each spectral classified water group, chla can be estimated with excellent accuracy (<10% errors). This approach would be essential when varieties of spectral water groups in nature considered.