A new method is presented to online detect the wear condition of a micro milling tool. The vibration signals of tested tools during the milling process are collected and analyzed in the time domain. Then, the length fractal dimensions of the tested tools are abstracted. Meanwhile, tools with different wear conditions are set as reference exemplars. Their clustering domain is obtained by collecting multistage time-domain signals of different exemplars and extracting their length fractal dimensions. Finally, the wear conditions of the tested tools are detected by comparing the length fractal dimensions of the tested tools with the clustering domain of the reference exemplars. The proposed method is experimentally verified based on the self-developed multifunctional micro machine. The clustering domain of seven reference exemplars is obtained, which includes the wear loss of micro milling cutter is 0, 5, 10, 15, 20, 45μm, and tipping. Then, the length fractal dimensions of 10 tested tools are extracted and compared with the clustering domain of the seven reference exemplars. The results show that the length fractal dimension of each tested cutter falls in the clustering domain that corresponds to the actual wear condition. Thus, it is effective and feasible to monitor micro milling tool wear based on length fractal dimension.