Intake of brassicaceous vegetables such as cabbage is associated with numerous health benefits. The major defense compounds in the Brassicales order are the amino acid-derived glucosinolates that have been associated with the health-promoting effects. This has primed a desire to build glucosinolate-producing microbial cell factories as a stable and reliable source. Here, we established-for the first time-production of the phenylalanine-derived benzylglucosinolate (BGLS) in Saccharomyces cerevisiae using two different engineering strategies: stable genome integration versus plasmid-based introduction of the biosynthetic genes. Although the plasmid-engineered strain showed a tendency to generate higher expression level of each gene (except CYP83B1) in the biosynthetic pathway, the genome-engineered strain produced 8.4-fold higher BGLS yield compared to the plasmid-engineered strain. Additionally, we optimized the genome-engineered strain by overexpressing the entry point genes CYP79A2 and CYP83B1, resulting in a 2-fold increase in BGLS production but also a 4.8-fold increase in the level of the last intermediate desulfo-benzylglucosinolate (dsBGLS). We applied several approaches to alleviate the metabolic bottleneck in the step where dsBGLS is converted to BGLS by sulfotransferase, SOT16 dependent on 39-phosphoadenosine-59-phosphosulfate (PAPS). BGLS production increased 1.7-fold by overexpressing SOT16 and 1.7-fold by introducing APS kinase, APK1, from Arabidopsis thaliana involved in the PAPS regeneration cycle. Modulating the endogenous sulfur assimilatory pathway through overexpression of MET3 and MET14 resulted in 2.4-fold to 12.81 mmol/L (=5.2 mg/L) for BGLS production. IMPORTANCE Intake of brassicaceous vegetables such as cabbage is associated with numerous health benefits. The major defense compounds in the Brassicales order are the amino acid-derived glucosinolates that have been associated with the health-promoting effects. This has primed a desire to build glucosinolate-producing microbial cell factories as a stable and reliable source. In this study, we engineered for the first time the production of phenylalanine-derived benzylglucosinolate in Saccharomyces cerevisiae with two engineering strategies: stable genome integration versus plasmid-based introduction of the biosynthetic genes. Although the plasmid-engineered strain generally showed higher expression level of each gene (except CYP83B1) in the biosynthetic pathway, the genome-engineered strain produced higher production level of benzylglucosinolate. Based on the genome-engineered strain, the benzylglucosinolate level was improved by optimization. Our study compared different approaches to engineer a multigene pathway for production of the plant natural product benzylglucosinolate. This may provide potential application in industrial biotechnology.