Drone delivery services are promising, but their application in takeaway delivery, along with the related economic and environmental impacts, remains underexplored. In this study, a multi-agent based simulation (MABS) model integrating drones, electric bicycles (e-bikes), and takeaway cabinets was designed to simulate the takeaway delivery process in a real scene and evaluate greenhouse gas (GHG) emissions and economic benefits under various delivery scenarios. The results indicated that during the use phase, the GHG emissions of drones were about 25 %-54 % higher than those of e-bikes, but from the perspective of life cycle, drone delivery showed certain environmental advantages when delivering 1 or 2 orders per flight, with GHG emissions 1 %-12 % lower than e-bike. Drone delivery demonstrated significantly higher efficiency and economics, with potential revenues 7-8 times that of e-bike delivery, and was expected to recover costs and achieve positive returns within two years. Takeaway cabinets, however, were major contributors to GHG emissions during the use phase, accounting for 50 %-60 % of total emissions, mainly due to the carbon intensity of electricity. Switching to clean energy sources could reduce GHG emissions by 60 %-70 %. The optimal collaborative delivery scenarios were affected by delivery distance, energy sources, and the number of orders per delivery.