One of the goals of Flexible Manufacturing System (FMS) is the mass production of personalized goods at cost comparable to the mass produced goods. This paradigm is referred to as mass personalization. To achieve this, the system has to seamlessly translate flexibility that can be achieved through the software that is responsible for the control of such system directly to the physical system, such that multiple distinct products can be produced in a non-batch mode. However, the present rigid design of Flexible Manufacturing Systems, which is characterized by static processing stations and rigid roll conveyor for part and material transportation, hampers this dream. In this paper, we propose a distributed architecture, which is implemented as Self-Organizing Flexible Manufacturing System (SoFMS), characterized by mobile processing stations that are capable of autonomously re-adjusting their location in real time on the shop floor to form an optimal layout depending on the mix of order inflow. This is achieved using the BEEPOST algorithm, an algorithm inspired by young honeybees’ collective behavior of aggregation in a temperature gradient field. An agent-based simulation paradigm is used to evaluate the viability and performance of the proposed system. The result of the simulation shows that processing stations are able to autonomously and optimally adjust their location depending on the mix of order inflow using the BEEPOST algorithm. This capability also results in higher throughput when compare to a similar system with static processing stations. This approach is expected to engender the capability for production of one-lot-size order in FMS, which is a requirement for mass-personalization.