Marine community composition is expected to be relatively stable in a natural environment over time but shift under increasing anthropogenic disturbances. In coastal waters, diatoms and dinoflagellates are two dominant phytoplankton functional groups. In this study, we developed an areal phytoplankton community composition index (APCI) that is based on the area of a scatter plot of dinoflagellate abundance (y-axis) vs diatom abundance (x-axis) using a time window of 1 year, 2 years or 3 years data. An APCI allows an ecological interpretation: it represents the fluctuability of a community composition within a time window and a temporal change between two neighbouring APCIs in a time series represents the stability of the composition. We used a 28-yr time series of monthly data on diatom and dinoflagellate abundance at four stations in Tolo Harbour and Channel (Tolo), Hong Kong to test the hypothesis that temporal changes in APCIs indicate environmental disturbances and to examine the applicability of APCI to indicate changes in nutrient conditions. We calculated the area (APCI) of a scatter plot of monthly data for 1-year, 2-year and 3-year windows, referred to as APCI-1y, −2y and -3y, respectively. The results show that, the fluctuability, is larger in APCI-3y than in APCI-1y, while the stability is stronger as temporal changes between neighbouring APCI-3y are smaller than between APCI-1ys. Temporal trends of APCIs are significantly correlated with those of dissolved inorganic nitrogen and phosphate concentration, which have declined after the implementation of a sewage diversion management plan in 1998. Hence, the APCI method is likely a robust indicator to assess a response of the phytoplankton community composition in a water body to environmental disturbances.