Understanding nitrogen dynamics in headwater streams is important for ascertaining how they influence downstream nutrient loads and identifying strategies for reducing loading through stream restoration. We compare nitrate uptake associated with two restoration approaches in headwater streams, Sheep Creek and Nunn Creek, of northern Colorado, USA. Segments of Sheep Creek were exclosed (fenced off) from open rangeland cattle grazing in the 1950s, allowing riparian corridors of these segments to naturally revegetate (passive approach), while other segments have been continually grazed. In 2003, restoration structures including cross vanes, J‐hook vanes, rootwads, log vanes, and bank riprap (structural approach) were installed along portions of Nunn Creek for trout habitat enhancement and local bank stabilization. We performed detailed physical characterizations and multiple nutrient injections of Br⁻ and NO₃ ⁻ to estimate transient storage and nitrate uptake in four reaches along Sheep Creek (two reaches exclosed from grazing and two grazed reaches) and two reaches along Nunn Creek (one with restoration structures and one without structures). Parameters of transient storage and nitrate uptake were estimated with the one‐dimensional transport with inflow and storage model run through universal inverse modelling code for optimization. Responses of transient storage and nitrate uptake to restoration techniques depended upon the type and extent of restoration implemented, as well the context and physical setting of each study reach. For example, in the higher‐gradient pair of Sheep Creek reaches, the restored reach showed greater nitrate uptake, while in the lower‐gradient pair of reaches, the non‐restored reach had greater uptake. At Nunn Creek, the reach with instream wood but without restoration structures exhibited more transient storage and nitrate uptake when compared with the reach with extensive J‐hook vane structures. Copyright © 2016 John Wiley & Sons, Ltd.