Oily sludge (OS) has long been regarded as a hazardous waste, and improper disposal may lead to serious environmental concerns and human health risks. Despite various methods having been proposed and applied to the treatment of OS, the oil occurrence states and properties in sludge are rarely characterized, which may directly link to the selection and effectiveness of treatment methods. Here, confocal laser scanning microscopy (CLSM), X-ray diffraction (XRD), gas chromatography (GC), and four components (SARA) analysis were utilized to characterize the changes in the oil occurrence states and compositions in OS samples before and after high-speed stirring (HSS) treatment. Our results show a substantial reduction in the oil concentration of OS after HSS treatment (from 32.98% to 1.65%), while SARA analysis reveals a similar oil composition before and after treatment, suggesting the broad applicability of HSS in removing oil and its insignificant selectivity towards various hydrocarbon components. This is further supported by the total petroleum hydrocarbon (TPH) analysis results, which show that the separated oil phase has a hydrocarbon composition similar to that of the original OS sample. The CLSM and fluorescence analysis suggest a homogeneous distribution of oil in the sludge, with relatively light components more concentrated in the pore systems between coarse mineral particles, whereas relatively heavy components tend to coexist with clay minerals. After HSS cleaning, both light and heavy components are removed to varying degrees, but light components are preferentially removed while heavy components tend to be retained in the sludge due to adsorption by clay minerals. This is consistent with TPH analysis, where a significant decrease in n-alkanes with lower carbon numbers (n-C14 to n-C20) was observed in the residual sample. Our findings demonstrate the dynamic response of oil occurrence states and compositions to the OS treatment process and highlight the importance of characterizing these fundamental properties prior to the selection of OS treatment methods.