Since 70% of the world's crops depend on pollinators for production and concerns are growing regarding insect decline, it is essential to implement robust and efficient monitoring of pollinator activity. However, traditional methods of pollinator monitoring are generally time-consuming and destructive. With the rise of technology, passive methods are being developed using computer vision or acoustic recording coupled with machine learning, and are offering the possibility to increase the temporal and spatial coverage of biodiversity and ecosystem functions monitoring. Passive acoustic monitoring is a promising method for tracking pollinators. However, it has rarely been implemented, and has mostly used relatively old machine learning methods. Deep learning methods, originally, developed for image analysis are beginning to be used for acoustic monitoring of various taxa, including flying insects. Here we proposed a method for quantifying pollinator activity in sunflower fields, based on the automatic identification of sounds produced by their beating wings. We tested a random forest and a deep learning algorithm on acoustic recordings using a new open access software dedicated to acoustic biodiversity monitoring, named TadariDeep. We found a higher performance of deep learning compared to random forest algorithms for the classification of pollinator flight sounds. The comparison of the acoustic monitoring of insects with pollinator activity estimated from a common protocol based on visual observations validates this method. We found that acoustic monitoring coupled with deep learning sound recognition provides a more realistic view of pollinator activity than visual observations, thanks to continuous monitoring. Acoustic monitoring of pollinators using deep learning, therefore, appears as a reliable method to quantify pollinator activity and might be used to monitor insect pollination over large spatial and temporal scales. Further improvements are however still needed for the species identification of pollinators.