In order to determine the phytoplankton community composition, the modification and optimizing of the CHEMical TAXonomy(CHEMTAX) running was carried out through samples grouping, successive run and evaluate the results for HPLC-pigment samples in the South China Sea(SCS). The vertical distribution of the ratio of pigment to total Chl a(TChl a) exhibited three different patterns, including increasing with depth pattern(e.g., But-Fuco), decreasing with depth pattern(e.g., Zea) and increasing at deep chlorophyll maximum(DCM) pattern(e.g., Hex-Fuco). The vertical profiles for Fuco/TChl a and Pras/TChl a was higher in coast than in the shelf and basin, and the Zea and Dv-Chl a expressed conversely. So the samples in the coastal stations must be separated for the cluster analysis group procedure in the SCS. Successive run was introduced into the CHEMTAX calculation and the output results were evaluated by the convergence of pigment/TChl a ratios. Most of the ratios were well converged at the fifth running, except Zea/TChl a for Prochlorococcus and Chl b/TChl a for prasinophytes and so on. To evaluate the fifth running’s results, haptophytes_8 and chlorophytes were two phytoplankton groups with much uncertainty. But the fifth estimated value was better than running once was supported by the regression evidence between the measured pigment concentration and calculation values. Synechococcus was another component with much mutability, and the CHEMTAX’s result should be compared to the flow cytometry’s cell abundance. The order for determining the phytoplankton community composition, the modification and optimizing of the CHEMical TAXonomy (CHEMTAX) running was carried out through the sample grouping, successive run and evaluate the results for HPLC-pigment samples in the South China Sea (SCS). The vertical distribution of the ratio of pigment to total Chl a (TChl a) exhibited three different patterns, including increasing with depth pattern (eg, But-Fuco), decreasing with depth pattern (eg, Zea) and increasing at deep chlorophyll maximum pattern (eg, Hex-Fuco). The vertical profiles for Fuco / TChl a and Pras / TChl a was higher in coast than in the shelf and basin, and the Zea and Dv-Chl a expressed conversely. So the samples in the coastal stations must be separated for the cluster analysis group procedure in the SCS. Successive run was introduced into the CHEMTAX calculation and the output results were were evaluated by the convergence of the pigment / TChl a ratios. Most of the ratios were well converged at the fifth running, except Zea / TChl a for Prochlorococcus and Chl b / TChl a for prasinophytes and so on. To evaluate the fifth running’s results, haptophytes_8 and chlorophytes were two phytoplankton groups with much uncertainty. But the fifth estimated value was better than running once was supported by the regression evidence between the measured pigment concentration and calculation values. Synechococcus was another component with much mutability, and the CHEMTAX’s result should be compared to the flow cytometry’s cell abundance.