Disordered multi-target scanning is an important application scenario of lidar target tracking technology. However, when dealing with scanning beam paths steered by cascaded prisms, it presents notable challenges, including highly non-linear and strongly coupled characteristics between the output vectors and prism rotation angles. Additionally, due to prismatic independence, there is no available coordinate system for visually observing the state of multiple prisms and the transitions between these states. In this work, we propose a rotation prism coordinate system, which offers coordinate conversion methods and mathematical expressions for distances. Two scanning modes, the fastest mode, and the shortest mode, along with two criteria, the total distance difference, and path curvature, are proposed for facilitating path selection and adhering to physical restrictions. To tackle the equivalent point-traveling salesman problem within this coordinate system, we employ an ant colony-taboo fusion algorithm. Our experiment results demonstrate a substantial improvement in efficiency, with up to a 2.77-fold enhancement compared to the original technique. This approach presents a promising solution for addressing the complexities of multi-target scanning with cascaded prisms in lidar target tracking. Future applications include environmental detection, rapid scanning and mapping, and identification of assembly parts in chaotic scenarios.