Underwater object classification is an attractive approach using acoustic remote sensing techniques due to its high coverage capabilities and limited costs as compared to manual method. In this approach Multi beam method is used for detection of underwater objects like sand, stone, silt etc. This under water object classification is performed by using model based approach which is achieved by implementing mathematical model of reservoir in Matlab using Finite Volume Method. An acoustic signal of particular frequency of Multi beam echo sounder is transmitted through this reservoir model. And thus from received signal strength various features are extracted, like backscatter strength, energy, power, surface roughness, skewness etc. It shows that these features are varying with changing seabed type, hence can be used to detect under water objects by using Multibeam echosounder of particular frequency. The scope of this project is found in underwater depth measurement using ultra sound to estimate the capacity, area elevation and sedimentation of a reservoir.