Chlorophyll-a (chl-a) concentration is considered to be the main measure of phytoplankton biomass. The location and intensity of the surface chl-a maximum in a coastal area are governed by daylight hours, air and seawater temperatures, and nutrient availability in the euphotic zone. The aim of this study is to model a back-propagation neural network (BP-ANN) for estimating chlorophyll-a concentrations from obtained input values. In this study an ANN structure of 3 input neurons and 1 output neuron is used. The 3 inputs represent sea surface temperature (SST), air temperature, and daylight hours, while the output represents chl-a concentration respectively and hidden layers number which is dependent to the application is determined as 20. The ANN structure, which is simulated in MATLAB, estimated the data of the experiments. When compared to current data, it can be said that these are successful results and they provide ANN for estimating chl-a. In our ANN approach, the effects of all input/output parameters can be evaluated and various outputs can be obtained for different environments and predicted maximum chl-a data.