Visual performances of honeybees have been extensively studied using free-flying individuals trained to choose visual stimuli paired with sucrose reward. By contrast, harnessed bees in the laboratory were not thought to be capable of learning a Pavlovian association between a visual stimulus (CS) and sucrose reward (US). For reasons as yet unknown, harnessed bees only learn visual cues in association with sucrose if their antennae are ablated. However, slow acquisition and low retention performances are obtained in this case. Here, we established a novel visual conditioning protocol, which allows studying visual learning and memory in intact harnessed bees in the laboratory. This protocol consists of conditioning the sting extension reflex (SER) by pairing a visual stimulus (CS+) with an electric shock punishment (US), and a different visual stimulus (CS-) with the absence of shock. Bees with intact antennae learned the discrimination between CS+ and CS- by using chromatic cues, achromatic cues or both. Antennae ablation was not only unnecessary for learning to occur but it even impaired visual SER conditioning because of a concomitant reduction of responsiveness to the electric shock. We thus established the first visual conditioning protocol on harnessed honeybees that does not require injuring the experimental subjects. This novel experimental approach opens new doors for accessing the neural correlates of visual learning and memory in honeybees.