In recent years, the DIDSON (Dual-frequency IDentification SONar),which can provide almost-video-quality images to identify objects even in turbid water has been used in enumerating fish populations, underwater structures inspection, oil/gas leakage detection and identification, underwater security, evidence searching, ship’s hulls and ports safety inspection, and underwater navigation and so on.In this paper, a designed vessel-mounted observing systems collected DIDSON data and GPS data according to designed lines in the Dishui Lake, shanghai city. Every line area was calculated according to the depth and sonar open-angle. Trained observers counted manually each fish image in every DIDSON file. Then based on the area and number of fish observed, the average density and zoning density of the Dishui lake was calculated. At last the whole fish number and distribution was calculated based on the ARCGIS software.The practice proved the method based on DIDSON Data is very feasible, effective and accurate for fishery resources estimating.