Transaminases inherently possess high enantiospecificity and are valuable tools for stereoselective synthesis of chiral amines in high yield from a ketone and a simple amino donor such as 2-propylamine. Most known ω-transaminases are (S)-selective and there is, therefore, a need of (R)-selective enzymes. We report the successful rational design of an (S)-selective ω-transaminase for reversed and improved enantioselectivity. Previously, engineering performed on this enzyme group was mainly based on directed evolution, with few exceptions. One reason for this is the current lack of 3D structures. We have explored the ω-transaminase from Chromobacterium violaceum and have used a homology modeling/rational design approach to create enzyme variants for which the activity was increased and the enantioselectivity reversed. This work led to the identification of key amino acid residues that control the activity and enantiomeric preference. To increase the enantiospecificity of the C. violaceum ω-transaminase, a possible single point mutation (W60C) in the active site was identified by homology modeling. By site-directed mutagenesis this enzyme variant was created and it displayed an E value improved up to 15-fold. In addition, to reverse the enantiomeric preference of the enzyme, two other point mutations (F88A/A231F) were identified. This double mutation created an enzyme variant, which displayed substrate dependent reversed enantiomeric preference with an E value shifted from 3.9 (S) to 63 (R) for 2-aminotetralin.