Abstract: A dynamic balance correcting apparatus comprising a learning unit and a balance confirmation unit for remeasuring the state of balance of a rotatable member which has been subjected to correction of its dynamic balance and for determining whether or not the rotatable member has been brought into a given balanced state. The results of the determinations by the balance confirmation unit are sequentially input to the learning unit. Then, the learning unit statistically calculates the ratio of success in the corrected state of dynamic balance of the rotatable member and automatically updates operational conditions of the apparatus in such a manner as to enable the ratio of success to reach the maximum value. It is therefore possible to correct dynamic unbalance of rotatable members without the need of experience and skill on the part of an operator, even when the ratio of success is reduced owing to time-dependent variations.