To forecast the population of brown planthopper (BPH), a major insect pest of rice in Mekong Delta in Viet Nam, a light trap network is used in the experiments where the BPH trapped density is considered as monitoring called BPH light trap surveillance network (BSNET). There are two problems in order to deploy the BSNET: the number of the light traps and their positions. In this paper, we propose a new approach to optimize the BSNET by determining the number of light traps needed and the position for every light trap node in the surveillance region based on HoneyCombarchitecture. The experiment results are performed on the Brown Planthoppers surveillance network for Mekong Delta in Viet Nam.