本文分别就纯方位角跟踪中的机动目标跟踪、有信号传输时延时的跟踪及一类特定的多目标跟踪问题进行了较为系统和深入的研究。首先，针对非机动目标提出一种智能距离参数化无味滤波方法，与传统方法相比，改进了跟踪初始性能、滤波精度以及优化了系统资源。其次，针对机动目标纯方位角跟踪提出一种将交互多模型和核粒子滤波结合的方法，在维持跟踪精度的前提下，大幅减少了所需粒子数，改善了系统实时性。-This paper bearings-only tracking of maneuvering target tracking, signal transmission time delay tracking and a special class of multiple target tracking problem systematically and deeply studied. Firstly, the maneuvering target puts forward an intelligent distance parameters of unscented filtering method, compared with the traditional method, improved the initial tracking performance, the filter accuracy and the optimization of system resources. Secondly, the maneuvering target bearings-only tracking presents an interactive multiple model and kernel particle filter binding method, in the maintenance of the tracking precision, greatly reduces the required number of particles, improves the system real time.