A microgrid control system using multi-agent technologies is proposed. In this control system, operation planning is realized based on generation and load forecasting by using artificial neural network and fuzzy systems. Unit commitment of generations includes start/stop of power generations and energy storages. Load following function is accomplished by an adaptive control system based on conventional PI control scheme. To include, in a flexible manner, new generations or loads into the microgrid, multi-agent technologies might be applied. In this paper, at first proposed Microgrid control system is outlined and then field tests outline is stated.