The poleward migration of the active regions’ magnetic flux on the solar surface plays an important role in the development of the large-scale field development, especially the polar-field reversal, which is a key process in the Babcock–Leighton-type solar dynamos. The poleward flux transport is nonuniform, centered around poleward surges as suggested by previous observations. The strong, long-lasting surges are related to activity complexes, and often result in violent polar-field reversal. However, the nonuniformity of poleward flux transport has not been evaluated quantitatively. We propose a statistical method to analyze the poleward flux transport during solar cycles 21–24 by considering the frequency distributions of the magnetic field at latitudes of poleward surges occurring during solar cycles. The nonuniformity is quantified as the kurtosis statistics representing the tailedness of the distributions. We test the method on results of surface flux transport simulations, and apply it to WSO, National Solar Observatory, MWO, and HMI data. We confirm that the poleward surges are of significance during solar cycles 21–24 in general. The kurtosis within a solar cycle is affected by different latitudes of the magnetic field and different data sources. The southern hemisphere of cycle 24 exhibits the largest kurtosis, agreeing with the super-surge concept from previous work. The significant nonuniformity of poleward flux transport originates from the nonrandomness of active regions, which favors the activity complexes as the origin of poleward surges.