A rapid and effective detection method is essential for water utilities to monitor variability of dissolved organic carbon (DOC) in source waters in order to apply strategies to minimize formation of disinfection by-products in treated waters. Ultraviolet absorbance at 254 nm (UVA254) and specific UVA254 (SUVA254) have been widely used as surrogates of concentration and reactivity of DOC, respectively. However, poor correlations between SUVA254 and specific trihalomethane formation potential (STHMFP) have been occasionally reported and the reliability of using SUVA254 to predict trihalomethane (THM) formation has been questioned. In this study, the correlations of SUVA254 and THM reactivity of three different DOC sources commonly found in water treatment facilities (aquatic carbon, soil carbon, and fecal matter) were evaluated. A 0.22 μm filter, instead of 0.45 μm filter, was used for water filtration to minimize the effects of colloidal materials on UVA254. UVA254 and DOC after chlorination were also examined and differential UVA254 and SUVA254 (ΔSUVA254) were compared to THM reactivity. Results showed correlations between UVA254 and DOC were source dependent suggesting natural humification and degradation processes did not alter DOC characteristics from its original sources. The STHMFP of river and soil DOC samples were comparable, whereas their UVA254 normalized THMFP were different (p<0.05), suggesting that UVA254 is a better indicator in predicting THM formation potential than DOC concentrations. ΔSUVA254 showed a stronger correlation with STHMFP than the conventional surrogate—SUVA254.