Metallic and non-metallic isomerases can be used to produce commercially important monosaccharides. To determine which category of isomerase is more suitable as a template for directed evolution to improve enzymes for galactose isomerization, L-arabinose isomerase from Escherichia coli (ECAI; E.C. 5.3.1.4) and tagatose-6-phosphate isomerase from Staphylococcus aureus (SATI; E.C. 5.3.1.26) were chosen as models of a metallic and non-metallic isomerase, respectively. Random mutations were introduced into the genes encoding ECAI and SATI at the same rate, resulting in the generation of 515 mutants of each isomerase. The isomerization activity of each of the mutants toward a non-natural substrate (galactose) was then measured. With an average mutation rate of 0.2 mutations/kb, 47.5% of the mutated ECAIs showed an increase in activity compared with wild-type ECAI, and the remaining 52.5% showed a decrease in activity. Among the mutated SATIs, 58.6% showed an increase in activity, whereas 41.4% showed a decrease in activity. Mutant clones showing a significant change in relative activity were sequenced and specific increases in activity were measured. The maximum increase in activity achieved by mutation of ECAI was 130%, and that for SATI was 190%. Based on these results, the characteristics of the different isomerases are discussed in terms of their usefulness for directed evolution of non-natural substrate isomerization.