Accurate estimation of regional terrestrial ecosystem carbon (C) balance is critical in formulating national and global adaptation and mitigation strategies in response to global changes. Since the regional C balance cannot be measured directly, it has been estimated using various models. In such studies, errors often exceeded the magnitude of the estimated C balance due to two types of uncertainties: noninclusion of some important factors in the C cycle and the fact that the C balance is a small difference between several large fluxes that can each be determined with only a limited accuracy. In this study, we propose new approaches to reduce these uncertainties and implement them in an Integrated Terrestrial Ecosystem C-budget model (InTEC). To minimize the first type of uncertainties, InTEC considers all the major factors presently known to affect C balance (including climate, atmospheric CO2 concentration, N deposition, and disturbances). To reduce the second type of uncertainties, InTEC estimates the C balance from historical changes in these factors, relative to the preindustrial period. InTEC is built on the basis of widely tested Century C cycling model, Farquhar's leaf photosynthesis model, and age-NPP relationships, and is constrained by N cycling. As a general regional-scale terrestrial ecosystem C budget model, InTEC has so far been applied to Canada's forests [Chen et al., this issue]. The sensitivity analysis showed that these two new approaches reduce the uncertainty in the C balance of Canada's forests substantially.