To develop an automatic software which can remotely classify the seabed sediment type using multibeam echosounder backscatter data, a supervised classification method is implemented, which is composed of both PCA (Principal Component Analysis) for classifying the extracted pixels from the backscatter data and RGA (Regional Growing Algorithm) for clustering the selected pixels. Although basic technologies, such as acoustic and ground truth data acquisition, data processing and classification strategy, were independently achieved, the classification accuracy needs to be increased using geoacoustic studies and sediment texture-based classification methods.