Summary In this work, a novel tomography algorithm with quantitative sensitivity-control capability is developed to overcome seismic data coverage imbalance, to improve the quality of tomography results, and to realize automatic target-oriented tomography. Unlike conventional tomography algorithms (a single optimization procedure), this new tomography algorithm cascades two optimizations - a sensitivity optimization and a data optimization. In this algorithm, we first design a targeted sensitivity distribution profile. Then, the sensitivity optimization problem is solved to automatically derive a nearly optimal data weighting scheme for sensitivity control. After that, this nearly optimal data weighting scheme is applied to the original tomographic-inversion problem to convert it into a sensitivity controlled tomographic-inversion problem (data optimization). Finally, by solving this sensitivity controlled tomographic-inversion problem, we can effectively circumvent the sensitivity imbalance issue that is very common in geophysical tomography applications. Furthermore, with this inversion-based sensitivity control technique, one is able to automatically implement the tomographic inversion in a target-oriented sense without introducing extra artificial discontinuities and inversion nonlinearity. Numerical examples validate this method and show its advantages over conventional tomography algorithms.