Protein engineering has been a research hotspot to improve the catalytic efficiency of industrially important enzymes. In the present study, a novel computational strategy was developed to in silico screen mutants with enhanced binding interaction between enzyme and substrate as well as catalytic efficiency. Through homology modeling and molecular dynamics (MD) simulation, four key residues related to substrate binding were identified in the endo-polygalacturonase BiPG28A from Bispora sp. MEY-1. Further analyses of the conformation, hydrogen bond interactions, and binding free energy revealed that lysine at position 129 (subsite - 2) has the strongest affinity to substrate. Biochemical and calorimetry experiments confirmed the functional role of Lys129 in substrate binding through non-covalent interactions. The common role of Lys129 was also verified in another GH28 endo-polygalacturonase. Distinguished from other protein engineering strategies involving structure resolution and construction of certain enzymes, this computational strategy represents an insightful and efficient approach to develop a "designed" enzyme with significantly enhanced binding affinity and catalytic efficiency.