Geographically Weighted Regression (GWR) is a new spatially local regression technology, which emerged in last decade and is attracting more and more attention in recent years. The technology is used for exploring spatial non-stationarity through embedding spatial locations in linear regression models. Compared with the traditional ordinary least squares regression method which uses global parameters, the GWR method uses local correlation coefficients to incorporate spatial heterogeneity which is usually non-stationary, thus more effectively dealing with spatial data. This paper first introduced the theoretical origin, principle, existing deficiencies and further expansion of GWR. At the meantime, in order to understand the research and application status of GWR more accurately, a literature survey was conducted. Then, the application situation of GWR in the soil and environmental sciences were reviewed and a look into the future was made. After years of development and practice, GWR has been proved to be a mature outstanding approach and should have a broad prospect of application in evaluation of resources and environment.