Aims Agromining aims to improve the fertility of naturally metal-rich soils by extracting metals, such as nickel (Ni), using hyperaccumulator plants. Ultramafic soils are characterized by low fertility levels, limiting hyperaccumulator yields. Here, we characterize the potential benefits for phytoextraction efficiency of co-cropping a Ni-hyperaccumulator (Odontarrhena chalcidica) and a legume (Vicia sativa), following a two-year field experiment. Methods A two-year field experiment was set up in an ultramafic zone in North-West Spain. Three treatments were tested: co-cropping, fertilized control with ammonium nitrate and non-fertilized control. Results Over the 2 years, co-cropping increased O. chalcidica’s biomass by 24% and 403% compared to fertilized and non-fertilized controls, respectively. Moreover, co-cropping had higher Ni-yields for both years, while fertilization had a negative effect on soil parameters. A non-metric multidimensional scaling analysis of the operational taxonomic units showed that the soil bacterial diversity changed over time. Soil exchangeable Ni and organic carbon influenced the phyla’s relative abundance. Metabolic genes were dominant and their relative abundances increased over time with co-cropping. Conclusion Pluriannual co-cropping of a hyperaccumulator with a legume improved both hyperaccumulator and Ni yields. In contrast, mineral fertilization was shown to be detrimental to some soil microbial parameters. Thus, ameliorating agromining by replacing mineral fertilizers would combine an eco-efficient strategy with sustainable metal recovery.