Site-specific herbicide spraying reduces herbicide use as it sprays only where weeds are detected. We studied the long-term impact of this weed-control measure on weed-impact indicators (crop yield loss, biodiversity, …). We developed a submodel to simulate the effects of site-specific spraying on weed floras and included this into the existing FlorSys model. The latter simulates multiannual multispecies weed dynamics and crop canopies at a daily time-step from cropping system, weather and soil. Global sensitivity and uncertainty analyses, based on 30-year-long simulations of different rotations and weather series, identified the most influential inputs and the most sensitive outputs. The cropping system (rotation with associated sowing patterns, herbicide products and treatment dates) was more influential than the spraying system (geometrical spraying pattern, weed detection). Finally, a real-life case study was simulated to demonstrate the feasibility of reconciling crop production with reduced herbicide use, thanks to site-specific spraying.