A central question in the stockman–animal relationship is how animals perceive humans depending on previous interactions with them. This study aimed at measuring the influence of a previous experience with humans on subsequent reactivity to humans of weaned piglets. Treatments differing in type (intrinsic vs extrinsic to the human) and valence (positive vs negative) of the reinforcements used over a 5-day standardised treatment period with a handler were delivered to 48 groups of three piglets following weaning: (1) gentle handling (GENHD), (2) food reward (FOOD), (3) gentle handling and food reward (FOODHD), (4) rough handling (ROUHD), (5) ball gun (e.g., plastic ball fired by a spring gun) without movement (GUN), (6) ball gun with movement (GUNHD), (7) passive human (PASSIVE, no reinforcement) and, (8) control (CONTROL, absence of additional experience with humans). The approach behaviour during a motionless handler phase was recorded weekly for 5 weeks after the treatment period. Thereafter, reactivity was scored (i.e., 0 = no escape to 4 = piglet escapes before handler makes any contact) during a handler approach phase. Following the treatment period, whereas piglets from all positive treatments approached the motionless handler, only those that received a gentle handling and were habituated to human contact accepted to be touched (average scores: 0.4, 0.3, 2.4 and 2.0 for FOODHD, GENHD, FOOD and PASSIVE, respectively, P < 0.0001). Rough handling was sufficient to induce fear and adding a mild physical stressor (i.e. ball gun shot) did not exacerbate the fear response (average scores > 3.8 for all negative treatments, P > 0.05). Unfamiliarity with the handler induced a natural vigilance and fear response since CONTROL piglets spent less time in contact with the handler than piglets from positive treatments until 2 weeks after the treatment period with the handler (28% vs 84%, 87%, 86% and 72% for CONTROL vs GENHD, FOODHD, FOOD and PASSIVE, respectively, P < 0.05) and they showed comparable reactivity scores to the handler approach with piglets from negative treatments until 5 weeks after the treatment (1.3, 2.7, 2.9 and 2.8 for CONTROL, ROUHD, GUN and GUNHD, respectively, P > 0.05). The perception of the handler was strongly modulated by the previous experience with her as behavioural differences between positive and negative treatments persisted for at least 5 weeks. For instance, during the last reactivity test, it is particularly noteworthy that all piglets from positive treatments touched the handler within 22 s, whereas 20% of piglets from negative treatments did not make contact (P < 0.0001). To conclude, piglets are able to associate humans with particular experiences and remember the past experience with humans during at least 5 weeks.