Strokes per minute (SPM) is one of the important parameters in beam pumping unit operation condition. Realtime accurate measurement of SPM is the basic principle to enhance the pump efficiency substantially. In view of the traditional methods by the formula calculating and sensor measuring, many problems of the additional measurement components, the high malfunction ratio, the difficulties of repairing and the accuracy influenced in the field operating conditions need to be considered. A softsensing approach for beam pumping unit SPM based on autocorrelation algorithm is proposed. The periodic correlated model is formulated among the SPM, load function and the driving motor input power function. The realtime input power is derived according to the driving motor input current and voltage. There are two kinds of features on the autocorrelation function. One is the periodic signal autocorrelation function period which is the same as input power function. The other is the noise signal autocorrelation function concentrated on the origin. On account of these two features, the input power function period can be identified accurately after interfering removed by the autocorrelation processing. Furthermore, the SPM is calculated according to the relationship between the SPM and the period of the electric power function The beam pumping unit SPM softsensing is realized accurately without additional peripherals. Simulation and an oilfield test results show that the proposed approach is practical and effective, high antiinterference, and the measurement error is less than 1%.