An ultrasound apparatus is described for externally treating kidney stone in human body. The apparatus has one or more ultrasound imaging transducers, a therapeutic ultrasound transducer, and a processing unit. Optimized delivering of ultrasound energy to the kidney stone from the therapeutic ultrasound transducer is based on real-time tracked state (e.g., position, movement shape, size, or combination thereof) of the kidney stone. The ultrasound imaging transducer(s) is configured to image the stone during the application of the therapy treatment. An optimization algorithm is implemented to control the therapeutic ultrasound transducer to apply different force vectors to the region of the stone. The effect of the vectors in the differing directions with respect to the stone may be detected and input to the optimization algorithm, which optimizes the therapy by adjusting one or more of the vectors. Therefore, less ultrasound energy may be delivered to break the kidney stone, and may reduce the possibility of damaging the surrounding tissues of the kidney stone.