β-Arbutin is a plant-derived glycoside and widely used in cosmetic and pharmaceutical industries because of its safe and effective skin-lightening property as well as anti-oxidant, anti-microbial, and anti-inflammatory activities. In recent years, microbial fermentation has become a highly promising method for the production of β-arbutin. However, this method suffers from low titer and low yield, which has become the bottleneck for its widely industrial application. In this study, we used β-arbutin to demonstrate methods for improving yields for industrial-scale production in Escherichia coli. First, the supply of precursors phosphoenolpyruvate and uridine diphosphate glucose was improved, leading to a 4.6-fold increase in β-arbutin production in shaking flasks. The engineered strain produced 36.12g/L β-arbutin with a yield of 0.11g/g glucose in a 3-L bioreactor. Next, based on the substrate and product's structural similarity, an endogenous O-acetyltransferase was identified as responsible for 6-O-acetylarbutin formation for the first time. Eliminating the formation of byproducts, including 6-O-acetylarbutin, tyrosine, and acetate, resulted in an engineered strain producing 43.79g/L β-arbutin with a yield of 0.22g/g glucose in fed-batch fermentation. Thus, the yield increased twofold by eliminating byproducts formation. To the best of our knowledge, this is the highest titer and yield of β-arbutin ever reported, paving the way for the industrial production of β-arbutin. This study demonstrated a systematic strategy to alleviate undesirable byproduct accumulation and improve the titer and yield of target products. KEY POINTS: •A systematic strategy to improve titer and yield was showed •Genes responsible for 6-O-acetylarbutin formation were firstly identified •43.79g/L β-arbutin was produced in bioreactor, which is the highest titer so far.