This paper presents a novel approach for multi-model 3D registration, which aims to simultaneously recover the motion of multiple objects in cluttered point clouds. The method is based on the Expectation-Maximization (EM) algorithm, which iteratively estimates the motion parameters of each object while handling outliers in the data. The approach does not require prior knowledge of the number of objects and can accommodate additional constraints. Theoretical analysis shows that the EM algorithm converges to the ground truth under certain conditions. The method is evaluated on both synthetic and real-world datasets, including table-top scenes and self-driving scenarios. The results demonstrate that the approach outperforms existing methods in terms of accuracy and robustness, especially when combined with state-of-the-art scene flow methods for dense correspondence estimation. The paper also provides a detailed theoretical analysis of the algorithm, showing that it can recover the true clusters under suitable initial conditions. The method is shown to be effective in both synthetic and real-world scenarios, and it is capable of handling a wide range of applications in robotics and computer vision.This paper presents a novel approach for multi-model 3D registration, which aims to simultaneously recover the motion of multiple objects in cluttered point clouds. The method is based on the Expectation-Maximization (EM) algorithm, which iteratively estimates the motion parameters of each object while handling outliers in the data. The approach does not require prior knowledge of the number of objects and can accommodate additional constraints. Theoretical analysis shows that the EM algorithm converges to the ground truth under certain conditions. The method is evaluated on both synthetic and real-world datasets, including table-top scenes and self-driving scenarios. The results demonstrate that the approach outperforms existing methods in terms of accuracy and robustness, especially when combined with state-of-the-art scene flow methods for dense correspondence estimation. The paper also provides a detailed theoretical analysis of the algorithm, showing that it can recover the true clusters under suitable initial conditions. The method is shown to be effective in both synthetic and real-world scenarios, and it is capable of handling a wide range of applications in robotics and computer vision.