Bone milling is one of the most widely used and high-risk procedures in various types of surgeries, and it is important to be noted that the experienced surgeon can perform such an operation safely. The objective of this article is to enhance the safety of the robot-assisted milling operation with the inspiration of human haptic perception. The emergence, coding and perception of the human haptic are introduced. Following this, a single axis accelerometer that measures the vibration of the surgical power tool is mounted in the robot arm, and the recorded acceleration signal is encoded as parallel stream of binary data. The data are subsequently inputted to the Hopfield network so as to identify the milling state. Inspired by human inference procedure, the fuzzy logic controller is introduced to control the robot to track the desired state when performing bone milling operations. A real-time implementation of the proposed method on a digital signal processing is also described. The experimental results in milling porcine spines prove that the robot accurately discriminates different milling states even when the additive noise is serious, and the safe motion control of the robot is also realized.