Abstract: Crop is one of the most important indicators to determine the farmland landscape pattern, crop productivity, and ecological benefits. It is necessary to clarify the changes in cropping patterns, thereby optimizing the farmland landscape during the sustainable intensification development of agriculture. Taking a typical county in Haihe Lowland Plain as the study area, this study aims to conduct the spatial distribution, variables of multiple cropping index, and cropping patterns at the county level from 2013 to 2019. A quadratic difference set was firstly used to collect the multiple cropping index, according to the smooth NDVI time-series curves by Harmonic Analysis of Time Series (HANTS). A Principal Component Analysis - Random Forest (PCA-RF) was then selected to classify the cropping patterns via the image interpretation using multi-temporal Landsat 8 OLI and Sentinel 2 images. General cropping patterns were interpreted including the cotton, pepper, peanut, maize, and tree crop single cropping, as well as the wheat-maize and tree crop double cropping. Finally, the transfer matrix was operated to quantify the transformations between every two cropping patterns, and then landscape metrics (e.g. Splitting Index (SPLIT), Shannon's Diversity Index (SHDI) and Shannon's Evenness Index (SHEI)) were calculated by Fragstats 4.2, further to describe the landscape fragmentation and diversity of the farmland. The overall accuracy and Kappa coefficient exceeded 90% and 0.84, respectively, after interpreting multiple cropping indexes using PCA-RF. Overall, the multiple cropping index increased from 163% to 174%, and the area of single cropping transformed to double cropping was 1.64 times that of the double cropping transformed to single cropping. The area of wheat-maize pattern remained stable, and the area of cotton single cropping decreased by 80.93%, with the area reduction of tree crop planting pattern increased by 64.54%. Cotton single cropping was the major pattern that transformed into wheat-maize double cropping, accounting for 81.15% of the transformed area. At the same time, the wheat-maize double and cotton single cropping were the main cropping patterns that transformed into the maize single cropping, accounting for 46.43% and 41.43% of the transformed area, respectively. The SHDI of the cropping pattern increased by 8.57%, and the SPLIT of wheat-maize double cropping and cotton single cropping increased significantly. In conclusion, the major variation in the farmland landscape was that: 1) The increase of the multiple indexes was mainly caused by transformation from cotton single cropping to wheat-maize double cropping. 2) The diversity of farmland landscape increased in two ways. One is that the cotton single cropping was transformed to other patterns (e.g. wheat-maize double cropping, tree crop double cropping, and cotton single cropping). The other is that more cropping patterns (such as peanut single cropping and maize single cropping) appeared, with the increased area of those cropping patterns in smaller areas. 3) Larger fragmentation of wheat-maize double cropping and cotton single cropping resulted in the more fragmented farmland landscape, in turn further preventing the large-scale crop production of the main cropping patterns.