Geological faults are essential requirements in geothermal system. Lineament is an important feature in showing subsurface structures as faults. In recent years automatic lineament extraction has gained huge popularity as it circumvents issues such as errors and time constraints that associated with manual extraction procedure. This study aims to identify subsurface lineaments automatically from gravity data. Gravity data from Olkaria geothermal field was measured, corrected and a Bouguer anomaly obtained. Polynomial regression and Quadratic Square trend surface analysis techniques were used to filter the data into regional and residual anomalies. Subsurface lineaments were generated from the residual contour map using Gaussian filter edge detection and curve extraction parameters of the line module of the PCI Geomatica. Density lineaments map was extracted which displayed faulting of the area. Orientation of lineaments was presented using rose diagrams where the main trend of the subsurface fault structures was observed to be dominantly in NW-SE and the predominantly in N-S. Area with high density lineaments were located around the center of the field, while the west area of the field had less fracturing