Cercospora leaf blight (CLB), primarily caused by Cercospora cf. flagellaris, is one of the most important diseases of soybean (Glycine max) in Louisiana. The pathogen produces cercosporin, a nonspecific toxin and an important virulence factor. There are no commercial cultivars with CLB resistance, and the pathogen has developed substantial resistance to the frequently used fungicides. Consequently, alternative methods are needed to manage CLB. One possibility is the RNA interference-based topical application of double-stranded (ds)RNA. The present study addressed the two most critical steps for this novel approach to be practical: inexpensively producing large quantities of dsRNA and identifying the right target genes for silencing. A screening method was developed to compare the effectiveness of Escherichia coli-produced dsRNAs targeting five fungal genes involved in cercosporin production for silencing in liquid culture. As much as 151.6 mg of dsRNA-containing total nucleic acids (TNAs) was produced from 1 liter of E. coli Luria broth culture using the L4440 vector. All tested dsRNAs reduced cercosporin production. However, significant target gene suppression was only detected in the cultures treated with dsRNAs from Avr4 and CTB8. The most potent dsRNA was from Avr4, which reduced 50% of cercosporin production at an estimated TNA concentration of 10.4 µg/ml (half maximal effective concentration [EC50]), and the least potent dsRNA was from HN-2, with an estimated EC50 of 46.7 µg/ml TNA. The present study paves the road for managing CLB under field conditions using dsRNA. Additionally, this approach could be adapted to identify the best dsRNAs to manage other fungal diseases.