We propose ``self-migrating threads'' as a new cluster computing paradigm for multi-agent applications, which can be viewed as the interactions among autonomous computing entities, each having its own objectives, behavior, and local information in a synthetic world. Self-migrating threads have both navigational autonomy of mobile agents and fine computation granularity of threads. They are also given the capability to construct system-wide logical networks, representing synthetic worlds. With those aspects, we expect that self-migrating threads provide multi-agent applications with good programmability and performance. We have designed the functionality of self-migrating threads and implemented a low-level migration library. In this paper, we discuss the feasibility of our design by considering the implementation techniques and basic migration performance.