This research focuses on modelling the relationship between wind conditions, sea roughness and GNSS reflections received from Low Earth Orbit (LEO). The motivation for this study lies in the recent development of a GNSS reflections receiver platform for the UK-DMC satellite and the numerous advantages proposed GNSS Reflectometry can provide in Earth Observation and global disaster monitoring. The fIrst part of the thesis focuses on the simulation procedure of received GPS-R Delay-Doppler Map (DDM). Airborne GPS-R scatterometric model has been adapted into this space-borne application research. Aft~r deriving DDM simulations according to reflection scenario, the results of two-dimensional data-model fItting are presented and analysed. The sensitivity discussion of current GPS-R model suggests some limitations of the modelling method, especially under medium and high wind speed ranges. In the second part, we investigate the inversion scheme of DDMs for the purpose of extracting a statistical wave model empirically. The similar model structure of DDM simulation is used but the processing order is turned over. After deconvolution, DDMs are inversed back to spatial energy maps and spatial slope probability maps. Three inversion algorithms are developed and compared. Preliminary synthetic and real data experiments give evidence of the feasibility of the inversion methodology. Finally, in the third part of this research, a new geometric wave slope statistical model is discussed in the context of wave fIeld simulations. The sensitivity of obtained statistical model is discussed in terms of wind speed, wave direction and observing incident angle. This provides an alternative view point to look into the wave slope probability properties and compensate the traditional theoretic and empirical wave modelling methods. Key words: GNSS-Reflectometry, Delay-Doppler Map inversion, wind conditions, sea surface roughness, slope probability density function, statistical wave slope model.