To improve the sensitivity, resolution, and other indicators of laser sensors, a mathematical model for optimizing laser sensor parameters based on data mining is constructed., Obtain dynamic parameters such as the moving distance and different angles of the target object within the measurement range of the laser sensor, and then use the EMD algorithm to mine the feature vectors of the dynamic data during the measurement of the laser sensor. The RBF neural network judges whether the current laser sensor parameters deviate from the theoretical values in the normal state based on the feature vectors, establishes a mathematical model for optimizing the laser sensor parameters, and optimizes the laser sensor parameters that deviate from the normal state. The experiment shows that this method can accurately distinguish whether there is a deviation in the distance between the laser sensor emitter and the object image reference, and can effectively optimize the parameters of the laser sensor. The sensitivity has been improved by 15.29%, and the application effect is good.