The compound 3-hydroxyphloretin is a typical dihydrochalcone that can be obtained in plants by the 3-hydroxylation of phloretin. Here, the flavin-dependent two-component monooxygenase (HpaBC) derived from Pseudomonas aeruginosa was used to convert phloretin into 3-hydroxyphloretin. Following molecular docking and sequence alignment, modifications to the substrate pocket and loop of Pa HpaBC were rationally designed, and mutant residues were selected. The results showed that the mutant Q212G/F292A/Q376N gave the best yield of 3-hydroxyphloretin and showed improved catalytic efficiency. Under optimal reaction condition, 2.03 g/L of 3-hydroxyphloretin was produced in the whole-cell catalysis experiment. Molecular docking and molecular dynamics simulations were used to analyze mutants and elucidate the potential mechanism. It was found that the increase in 3-hydroxyphloretin yield was due to the improvement in the flexibility of the loop and the expansion of its substrate pocket. This strategy based on loop and substrate pocket modification has significance in the engineering of Pa HpaB.