The mosquito Aedes aegypti is the major vector of arboviruses like dengue, Zika and chikungunya viruses. Attempts to reduce arboviruses emergence focusing on Ae. aegypti control has proven challenging due to the increase of insecticide resistances. An emerging strategy which consists of releasing Ae. aegypti artificially infected with Wolbachia in natural mosquito populations is currently being developed. The monitoring of Wolbachia-positive Ae. aegypti in the field is performed in order to ensure the program effectiveness. Here, the reliability of the Matrix‑Assisted Laser Desorption Ionization‑Time Of Flight (MALDI‑TOF) coupled with the machine learning methods like Convolutional Neural Network (CNN) to detect Wolbachia in field Ae. aegypti was assessed for the first time. For this purpose, laboratory reared and field Ae. aegypti were analyzed. The results showed that the CNN recognized Ae. aegypti spectral patterns associated with Wolbachia-infection. The MALDI-TOF coupled with the CNN (sensitivity = 93%, specificity = 99%, accuracy = 97%) was more efficient than the loop-mediated isothermal amplification (LAMP), and as efficient as qPCR for Wolbachia detection. It therefore represents an interesting method to evaluate the prevalence of Wolbachia in field Ae. aegypti mosquitoes.