Abstract: Productivity of cultivated land needs to be greatly improved for the ever-growing demand for food security in China. Monitoring the productivity of cultivated land can contribute to figuring out the background of food security for the decision-making on protection policies of targeted farmland. In this study, a Savitzky-Golay filter was selected to rebuild the smoothing Enhanced Vegetation Index (EVI) curve of crop growth using the moderate resolution imaging spectroradiometer (MODIS)-EVI data. The moving window method was utilized to eliminate the possible impact of climate factors on crop yield and extract the Multiple Crop Index (MCI) of cultivated land from the EVI growth curve by the quadratic difference method. The threshold method was used to calculate the optimal growth time (Most Active Day, MAD) and characterize the change of farmland productivity. Taking Sichuan Province of China from 2001 to 2017 as a research object, an attempt was made to analyze the spatial distribution and potential characteristics of cultivated land productivity. The results were obtained: 1) The Multiple Cropping Index (MCI) of 93.5% of cultivated land in Sichuan Province remained unchanged from 2001 to 2017. The areas with changed MCI were mainly concentrated in Chengdu city and all around, such as Meishan and Deyang. The cropping system in the MCI constant zone was dominated by a single-cropping system (accounting for 88.3% of the total arable land) and a double-cropping system (2.7% of that); 2) There was a great variation in the productivity of cultivated land in Sichuan Province from 2001 to 2017. The area that single-season crop productivity significantly reduced and increased accounted for 24.0% and 33.2% of the total cultivated land. The decreased areas were distributed mainly in the northeastern (Nanchong, Guangyuan) and southern (Neijiang, Zigong, Yibin) Sichuan. In the double-season crop, the areas with the changing productivity of the first and second crops were 62.1% and 49.4% of cultivated land, respectively, indicating a trend of first decreasing and then rising and fluctuating declining. 3) There was much potential growth for the double-season crops, but not for the single-season crops. The cultivated land, whose productivity improvement potential of single-season crops was less than 40%, accounted for 83.8% and 71.4% in the productivity upgrading and declining areas, respectively. Those in productivity improvement potential of greater than 40% for the first and second crops in a double-cropping system were 89.5% and 67.6%, in productivity decreased areas, respectively. Over 50% of the areas indicated productivity improvement potential greater than 80% in the cultivated land of declining productivity. The areas of high productivity improvement potential were located mainly in the Chengdu Plain and surroundings, which were dominated by the double cropping system. The proposed method can be expected to monitor the large-scale and long-term sequence of productivity change, thereby identifying the productivity improvement potential of farmland. The finding can provide a sound reference to optimize farmland productivity in the protection policy and planning.