Abstract: Based on multi-sensor data fusion a method of analysis is presental that fuses together original statistical data obtained from means and variances of multi-channel signals and indentify the cutting tool stain by means of their higher-order terms and artificial neural network and fault tree theory.Simulations show that the method is quite effective in identifying different cutting-tool wear levels.Experiments of monitoring boring breakage on FMC show that it is a new practival and feasible method to monitor cutting-tool states by the use of multi-sensor data sampled from a new type flow acoustic emission sensor and accelerator sensor.