The accurate identification of surface water pollution sources is very important for formulating remedial schemes and determining the parties responsible for sudden water pollution incidents. According to the basic principles and theories of pollution source identification, an overview of the existing methods is presented in this study, including mathematical simulation and footprint analysis methods. The mathematical simulation methods comprise direct solution methods (analytical and regularization-based methods) and indirect solution methods (simulation optimization methods, probability-based statistical methods, and coupling algorithms). The uncertainty of pollution source identification is discussed in terms of the pollution source information, observational data, surface water model, hydrodynamic conditions, and pollutant properties. The uncertainty and timeliness of surface water pollution source identification are the key factors that affect the valid application of these methods, and thus further research is needed in areas such as method coupling, improving the identification efficiency, uncertainty analysis, and determining the pollutant properties.