Symbiotic mutualisms provide critical ecosystem services throughout the world. Anthropogenic stressors, however, may disrupt mutualistic interactions and impact ecosystem health. The plant‐rhizobia symbiosis promotes plant growth and contributes to the nitrogen (N) cycle. While off‐target herbicide exposure is recognized as a significant stressor impacting wild plants, we lack knowledge about how it affects the symbiotic relationship between plants and rhizobia. Moreover, we do not know whether the impact of herbicide exposure on symbiotic traits or plant fitness might be ameliorated by plant or rhizobial genetic variation. To address these gaps, we conducted a greenhouse study where we grew 17 full‐sibling genetic families of red clover (Trifolium pratense) either alone (uninoculated) or in symbiosis with one of two genetic strains of rhizobia (Rhizobium leguminosarum) and exposed them to a concentration of the herbicide dicamba that simulated “drift” (i.e., off‐target atmospheric movement) or a control solution. We recorded responses in immediate vegetative injury, key features of the plant‐rhizobia mutualism (nodule number, nodule size, and N fixation), mutualism outcomes, and plant fitness (biomass). In general, we found that rhizobial variation more than plant variation determined outcomes of mutualism and plant fitness in response to herbicide exposure. Herbicide damage response depended on plant family, but also whether plants were inoculated with rhizobia and if so, with which strain. Rhizobial strain variation determined nodule number and size, but this was herbicide treatment‐dependent. In contrast, strain and herbicide treatment independently impacted symbiotic N fixation. And while herbicide exposure significantly reduced plant fitness, this effect depended on inoculation state. Furthermore, the differential fitness benefits that the two rhizobial strains provided plants seemed to diminish under herbicidal conditions. Altogether, these findings suggest that exposure to low levels of herbicide impact key components of the plant‐rhizobia mutualism as well as plant fitness, but genetic variation in the partners determines the magnitude and/or direction of these effects. In particular, our results highlight a strong role of rhizobial strain identity in driving both symbiotic and plant growth responses to herbicide stress.