Optical network unit (ONU) placement is an essential challenge in fiber-wireless (FiWi) access networks, which affects the development cost and quality of services. Recently, some intelligent methods for ONU placement were developed and achieved encouraging results; however, their performance is far from the ideal state. A hybrid ONU placement method (AOFCM) based on Fuzzy C-Means (FCM) and Arithmetic Optimization (AO) algorithm is presented in this work, which makes full use of the merits of both algorithms. The AOFCM is a population-based iterative algorithm that models the ONU placement as a multi-modal optimization problem. It starts the optimization process with a collection of individuals. An individual contains the coordinates of multiple ONUs. AOFCM sequentially applies the FCM and the AO algorithms to the population such that the fitness of each individual is improved. With a test on four placement benchmarks, the proposed AOFCM is proven to significantly improve counterpart algorithms in cost minimization and convergence speed.