The gas puff imaging (GPI) diagnostic has been developed and operated on EAST since 2012, and the time-delay estimation (TDE) method is used to derive the propagation velocity of fluctuations from the two-dimensional GPI data. However, the TDE method is difficult to analyze the data with fast transient events, such as edge-localized mode (ELM). In consequence, a method called the spatial displacement estimation (SDE) algorithm is developed to estimate the turbulence velocity with high temporal resolution. Based on the SDE algorithm, we make some improvements, including an adaptive median filter and super-resolution technology. After the development of the algorithm, a straight-line movement and a curved-line movement are used to test the accuracy of the algorithm, and the calculated speed agrees well with preset speed. This SDE algorithm is applied to the EAST GPI data analysis, and the derived propagation velocity of turbulence is consistent with that from the TDE method, but with much higher temporal resolution.