Changes in soil and vegetation management can impact strongly on the rates of carbon (C) accumulation and loss in soil, even over short periods of time. Detecting the effects of such changes in accumulation and loss rates on the amount of C stored in soil presents many challenges. Consideration of the temporal and spatial heterogeneity of soil properties, general environmental conditions, and management history is essential when designing methods for monitoring and projecting changes in soil C stocks. Several approaches and tools will be required to develop reliable estimates of changes in soil C at scales ranging from the individual experimental plot to whole regional and national inventories. In this paper we present an overview of soil properties and processes that must be considered. We classify the methods for determining soil C changes as direct or indirect. Direct methods include field and laboratory measurements of total C, various physical and chemical fractions, and C isotopes. A promising direct method is eddy covariance measurement of CO2 fluxes. Indirect methods include simple and stratified accounting, use of environmental and topographic relationships, and modeling approaches. We present a conceptual plan for monitoring soil C changes at regional scales that can be readily implemented. Finally, we anticipate significant improvements in soil C monitoring with the advent of instruments capable of direct and precise measurements in the field as well as methods for interpreting and extrapolating spatial and temporal information.