The invention discloses an intelligent automobile autonomous tracking path planning method, which comprises the following steps of (1) initializing the value of an intelligent automobile deviating from a rail center position; normalizing the value to a position between -1 and 1, wherein the plus or minus sign represents the deviating direction of the intelligent automobile; (2) listing and writing a P value required by a PID controller according to the normalized value in accordance with a formula that P=a^3/sc+a^2/ec+a/yc+P[min], wherein sc is a coefficient of a cubic term; ec is a coefficient of a quadratic term; yc is a coefficient of a first order term; P[min] is the minimum basic P value; (3) detecting the value of the intelligent automobile deviating from the current operation rail center position; performing path planning on the intelligent automobile by the PID controller of the intelligent automobile according to the value. By using the method, the intelligent automobile can better operate on the operation rail; if the operation effect is poor, the optimum state can be reached only through properly debugging the value of each parameter in the cubic function in the P value (proportion value).