Abstarct: Precision navigation in an effective way has become an important aspect of many applications such as military and commercial applications. Currently, GPS stands as the preferred navigation system to the wide variety of users and has provided an unprecedented sub-meter level of accuracy, but whose signals are vulnerable to interference and jamming. There are various approaches to deter-mine the position and navigation in a GPS denied environment in the literature including, but are not limited, Inertial Navigation System (INS), Terrain Refer-enced Navigation (TRN) and so on. Nowadaysanother method of navigation that has been given a large amount of attention is by use of a device known as Gravity Gradient Instrument (GGI) devices which measure the spatial derivatives of gravi-ty and Nowadays have experienced remarkable improves in accuracy. This Thesis proposes two methods to navigate using Earth’s gravity field data. The first is a novel and robust method for position determination by using Neuro-fuzzy modelling and GGI’s Data which doesn’t have aforementioned weaknesses. The parameters of Generated TSK network is optimised by a new hybrid learning algorithmwhich uses Modified Shuffled Frof Leaping Algorithm (MSFLA) and Least Square Stimator (LSE).The obtained acuuracy in smooth terrain is about 16.61 meters, which reveals that proposed method have a good potential for position determination in smooth terrain . Also the suitability of proposed method for use in different altitudes is compared. At the end of this paper current limitations of this method areidentified. The second method uses a linear kalman filter to fuse INS navigation data and gravitational navigation system. This system is suitable for Vehicles with long flight duration. At defult condition we can achieve the accuracy of 45.92 meters and 25.35 meters on Rough and Smooth terrain.