The invention relates to a multi-model soft measurement method for pumping efficiency of an oil pumping well on the basis of data. The method comprises the following steps of creating a soft measurement system for pumping efficiency and separately collecting an upper-stroke average load, a lower-stroke average load, wellhead pressure, the stroke and stroke frequency of an oil pump, wellhead liquiddischarge amount, upper-stroke average current data of a motor and lower-stroke average current data of the motor; training and establishing multiple pumping-well pumping-efficiency soft-measurementmodels according to historical production data; substituting the newly collected data into the trained models to obtain pumping efficiency data in order to achieve soft measurement of the pumping efficiency of the oil pumping well. The method can achieve online continuous measurement of production data, is low in measurement cost and great in real-time property and can avoid the problems of data loss, recording errors and the like caused by manual recording. The method can also avoid the problem that a decision-making process falls behind an actual working condition of the oil pumping well dueto the fact that oil liquid pumped from the wellhead is subjected to sampling assay. The prediction accuracy is high, and the problem that the prediction precision is influenced since collected datasamples contain noise can be effectively handled.