Subsurface cavities are very susceptible subsurface locations for down-lifting of a dam construction.In order to detect the low-density zones of a power plant a micro-gravity survey was conducted in a site considered for construction of a power plant site in Iran.First we gain the residual anomalies through bouger anomalies and then we design an Artificial Neural Network(ANN)which is trained by a set of training data.The ANN was tested for both synthetic and real data.For real data some suitable features are derivate from residual anomalies and applied to the Subsurface cavities are very susceptible subsurface locations for down-lifting of a dam construction. In order to detect the low-density zones of a power plant a micro-gravity survey was conducted in a site considered for construction of a power plant site in Iran. First we gain the residual anomalies through bouger anomalies and then we design an Artificial Neural Network (ANN) which is trained by a set of training data. The ANN was tested for both synthetic and real data. For real data some suitable features are derivate from residual anomalies and applied to the