Weed populations in arable fields are often spatially and temporally heterogeneous, and high-density weed patches require intensive weed management. In areas with no or few weeds, input for weed control can be reduced. This chapter describes manual and geographic information system–based weed mapping approaches and sensor-based weed identification technologies and their application to create weed maps. It also describes the adoption of patch sprayers based on application maps and real-time sensor information. This chapter shows that site-specific weed control in cereals, oilseed rape, sugar beet and maize using Global Navigation Satellite System-controlled patch sprayers can reduce herbicide input by 20–60%. It looks ahead to the development of precision farming technologies for mechanical weeding, including automatic steering systems for intra-row hoeing and robotic inter-row weeding.