Enhanced sequestration of plant-carbon (C) inputs to soil may mitigate rising atmospheric carbon dioxide (CO2) concentrations and related climate change but how this sequestration will respond to anthropogenic nitrogen (N) and phosphorous (P) deposition is uncertain. We couple isotope, soil C fractionation and mesocosm techniques to assess the sequestration of plant-C inputs, and their partitioning into C pools with different sink potentials, under an experimental gradient of N and P deposition (0, 10, 30, 60 and 100 kg N ha−1 yr−1; and 0, 2, 6, 12 and 20 kg P ha−1 yr−1). We hypothesized that N deposition would increase sequestration, with the majority of the C being sequestered in faster cycling soil pools because N deposition has been shown to accelerate the turnover of these pools while decelerating the turnover of slower cycling pools. In contrast to this hypothesis, sequestration into all soil C pools peaked at intermediate levels of N deposition. Given that P amendment has been shown to cause a net loss of soil C, we postulated that P deposition would decrease sequestration. This expectation was not supported by our data, with sequestration generally being greater under P deposition. When soils were amended simultaneously with N and P, neither the shape of the sequestration relationship across the deposition gradient, nor the observed sequestration at the majority of the deposition rates, was statistically predictable from the effects of N and P in isolation. The profound nonlinearities we observed, both for total sequestration responses and the partitioning of C into soil pools with different sink potentials, suggests that the rates of N and P deposition to ecosystems will be the critical determinant of whether they enhance or decrease the long-term sequestration of fresh plant-C inputs to soils.