Solid surface excitation-emission matrix (EEM) fluorescence spectroscopy, a leading technique, is used to characterize structural composition of solid organic matter. The EEM spectra were directly measured using solid samples, instead of extraction of dissolved organic matter. Hence the technique has feasibility, practicality and large amount of information. In this study, the solid surface EEM spectroscopy coupled with parallel factor analysis (PARAFAC), hierarchical cluster analysis (HCA), and classification and regression tree (CART) was applied to extract fluorescent components of the surface sediments in Wuliangsuhai Lake, to track potential factors of the organic matter, and to reveal spatial variations of the components. The ten samples were collected along a pollution gradient. The solid surface EEM fluorescence spectroscopy were measured using the untreated sample and the thermally treated samples, and the spectroscopy of the organic matter were obtained by the difference between the former and the latter. Four fluorescence components (C1 to C4) were extracted by the PARAFAC. The C1 was associated with the tryptophan-like material, which could be derived from in-situ source. The C2 was relative to the fulvic-like material, and the C3 and C4 were concerned with the Vis humic-like and UV humic-like materials respectively, which could be derived from ex-situ source. The total abundances of the four components were the highest in the northern region, followed by the southern and the central regions. The abundances of the C2 to C4 were higher than those of the C1, suggesting that the organic matter was mainly derived from the ex-situ source. The C1 in the southern region was relatively higher than that in northern and central, indicating that the C1 could be relative to the metabolism of many more aquatic plants. The decreasing order of the C2 was northern>southern>central. The C3 was higher in the northern region than those in the southern and central regions, so the C3 was the representative of the sediment in the north. The trend of the C4 was similar to the C2. Based on the fluorescence component HCA, the C2 and C4 might have the terrestrial source, while the C1 and C3 were the potential factors of the organic matter in the sediments. Based on the sample HCA, the sampling sites were grouped into three clusters, i. e. the high-pollution (the northern region), the medium-pollution (the southern region) and the low-pollution (the central region). With the CART model, the C3 and C1 were verified as the potential indicators, which provided clearer classification of the sediments, and offered technical support forsubsequent exploration of pollution characteristics and sources.