Surveillance of malaria vectors in Africa is most often accomplished using CDC-type light traps or human landing catches (HLCs). Over the past 30 yr, a variety of commercial and experimental mosquito traps have been developed for residential mosquito control or for improved surveillance of disease vector species, including the BG Sentinel (BGS) trap. To optimize collection of Anopheles gambiae Giles using this trap, BGS traps were modified with an opening (vent) added to the trap base to decrease exhaust airflow. Four traps configurations were tested with colony-reared host-seeking female An. gambiae in free-flying laboratory enclosures. Six attractant treatments (three attractants: BG-Lure, Limburger cheese, and a blank, with and without CO₂) were tested concurrently. Across all trap—attractant combinations, significantly more mosquitoes (P < 0.05) were collected in standard, unmodified traps set upright (mean: 10.0) or upside down (9.8) than vented traps, whether upright (5.9) or upside down (7.0). The BG-Lure CO₂ and Limburger Cheese CO₂ bait combinations were more attractive than the other attractants tested alone. All attractant combinations collected significantly more mosquitoes than unbaited traps. Field studies are needed to determine if BG-Lure CO₂- or Limburger cheese CO₂-baited BGS traps are comparable with HLCs in collecting host-seeking An. gambiae.