The invention discloses a method for evaluating the validity of SCADA data on the basis of amount of reactive unbalance. The method is characterized by acquiring the amount of reactive unbalance of both ends of a line by computing according to a voltage amplitude on both ends of the line, a current amplitude of the line, the impedance of the line, and a susceptance-to-ground parameter of the line; and acquiring the amount of reactive unbalance according to the reactive power of SCADA measurement data at both ends. In an ideal case, the computed amount of reactive unbalance is approximately equal to the measured amount of reactive unbalance. If a difference value between the computed amount of reactive unbalance and the measured amount of reactive unbalance is very large, the validity of the SCADA data at that time segment is very low. The method has beneficial effects of using the reactive power of measurement data as an evaluation index to accurately evaluate the validity of the measurement data, thereby accurately finding out data with poor validity in the measurement data and fast removing the data with poor validity.