The invention relates to the field of mobile robot track tracing, in particular to a robot track tracing method based on a Skinner operant conditioning automata. According to the robot track tracing method, firstly, an operation and state set of a robot is established, a possibility set from corresponding states to operations is established, and the sets are uniformly distributed; secondly, one operation is selected randomly, the corresponding position change is calculated, an orientation function is calculated according to a distance from a target track, action possibility distribution is adjusted based on an operant conditioning theory according to the orientation function, and a system entropy is calculated; thirdly, when the system entropy is close to a minimum value, studying is finished. At the moment, a possibility matrix is optimal. According to the robot track tracing method, operant conditioning behaviors of humans and animals can be well simulated, the intelligent level of robots can be improved to endow the robots with relatively strong abilities of self-learning, self-organization and self-adaption, self-control on condition parameters can be realized, and track tracing can be successfully carried out.