Studies have been undertaken to explore the applicability of different kinetic models for the performance appraisal of upflow anaerobic sludge blanket (UASB) reactors treating wastewater in the range of 300-4000 mg COD/l. Three kinetic models namely, Monod, Grau second-order, and Haldane model are considered for the analysis. Both linear and nonlinear regressions have been performed to examine the best-fit among the kinetic models. In this process, five error analysis methods have been used to analyze the data. Apart from optimization of kinetic coefficients with minimization of associated errors, prediction of effluent COD has also been undertaken to verify the applicability of kinetic models. In both the cases, Grau second-order model is found to be the best class of fit for wide range of data sets in UASB reactor.