Good protein thermostability is very important for the protein application. In this report, we propose a strategy which contained a prediction method to select residues related to protein thermal stability, but not related to protein function, and an experiment method to screen the mutants with enhanced thermostability. The prediction strategy was based on the calculated site evolutionary entropy and unfolding free energy difference between the mutant and wild-type (WT) methyl parathion hydrolase enzyme from Ochrobactrum sp. M231 [Ochr-methyl parathion hydrolase (MPH)]. As a result, seven amino acid sites within Ochr-MPH were selected and used to construct seven saturation mutagenesis libraries. The results of screening these libraries indicated that six sites could result in mutated enzymes exhibiting better thermal stability than the WT enzyme. A stepwise evolutionary approach was designed to combine these selected mutants and a mutant with four point mutations (S274Q/T183E/K197L/S192M) was selected. The T (m) and T (50) of the mutant enzyme were 11.7 and 10.2 A degrees C higher, respectively, than that of the WT enzyme. The success of this design methodology for Ochr-MPH suggests that it was an efficient strategy for enhancing protein thermostability and suitable for protein engineering.