Ground Penetrating Radar (GPR) is an effective tool to measure the geological properties. A lot of information can be interpreted from the GPR data, such as soil water content. One of the common approaches is to determine the apparent electrical permittivity from the transmission velocity of the impulse electromagnetic wave, and to use empirical relationships to estimate the soil water content. For example, Ferre equation & Topp equation are all expressing the relationship between soil water content and electrical permittivity. However, this method has some limitations; most notably the necessity to determine the velocity from a known depth to a reflecting surface. Therefore, another approach using the frequency dependent attenuation represented by a parameter called Q* was tested and studied in this thesis. The Q* method was evaluated using laboratory measurements, which consists of a series of experiments. A new empirical model was established using experiments where Q* was estimated from measurements on a soil sample with known water contents using two types of antennas (1.6 GHz & 2.3 GHz). Finally, the adaptability of Topp equation and Ferre equation were verified, and a new empirical equation was defined. What’s more, the other method using Q* was proved to be feasible.