The success of wildflower strips for conserving pollinator biodiversity is often hampered by their poor sustainability, attributable to the short duration lifespan of the sown species. This three-year experiment aimed to: i) select the plant species that survived the agronomic disturbance practices implemented and ii) verify which crop management approach favoured their sustainability. Six experimental strips along the longest edges of the adjacent wheat crop were sown during the fall of 2019. The annual wildflower species that showed the best performances in emergence dynamics and seedling growth were some wildflowers derived from segetal weeds that are presently rare in conventional agroecosystems. The species Centaurea cyanus, Agrostemma githago, Glebionis coronaria among others attained the phenological stage of flowering most consistently, and also had the lowest mortality rates in the plant community studied. Despite preparing a stale seedbed, weeds were the most significant obstacle to the sustainability of the strips over time. Soil harrowing at the end of the summer lifecycle led to better plant survival performances (10.9 %) compared to senescent plant shredding (4.8 %). Harrowing also resulted in a greater wildflower survival the following year, as well as a higher number of pollinator visits. Honeybee visits were decreased by wildflower strip thinning over time, probably due to their typical constancy in the daily foraging choice for the same abundant species. A similar reduction was observed by the Lepidoptera. In contrast, generalist pollinators (i.e. Syrphidae, Bombyliidae, solitary bees and Coleoptera) were the least demanding pollinators in terms of the plant biodiversity of the sustainable wildflower strip. Harrowing led to a greater biodiversity of both wildflowers and pollinators (Shannon index, H′), and a lower weed dominance (Simpson index, D), compared to shredding. In summary, some segetal wildflowers could be incorporated into sustainable wildflower strips as they are self-seeding.