The invention relates to a power grid parameter estimation method considering probability distribution, which belongs to the field of power system dispatching automation. The power grid parameter estimation method comprises the steps of: acquiring latest model data of a power grid at a period of 15 minutes, carrying out power grid parameter identification and estimating an equipment parameter of the power grid; substituting the parameter estimation result under a section into power grid calculating parameters, and recompleting state estimation calculation; calculating a number difference value of large measuring error points before and after substituting parameter estimated values, and saving the result; acquiring a reliable parameter estimation result of one equipment the same day by taking time as an index, and calculating a parameter estimation daily available rate of the equipment; setting 0-2 times of an original value of the equipment as a statistical interval, and counting probability distribution and an average large measuring error point difference value of the estimated value in each interval at the set interval range (defaulted as base value*5%); and calculating a probability optimization target of parameter values, and setting the equipment parameter estimated value corresponding to the maximum value of the probability optimization target to be an equipment daily parameter estimated value. The power grid parameter estimation method increases the power grid equipment parameter estimation accuracy.