In order to accurately understand the intensive use level, the spatial difference characteristics of farmland and the influencing factors among different regions, this paper applied the multi-factor comprehensive evaluation and the K-mean clustering methods and the spatial measurement model to study the change of intensive use degree of farmland and to explore the internal factors affecting the level of intensive use of farmland in the four eastern cities of Gansu Province. Results show that: 1) in 2016, the intensive use level of farmland in the research area was generally at a relatively coarse utilization level (mean value were 0.377 2) with obvious regional differences. Grades III and IV were located in the northeast and southwest of the study area, and I and II were located in the middle of the study area. So a“dumbbell”shape was formed from regional perspective; 2) the global Moran index was 0.142 2, indicating that the intensive utilization degree of farmland in each county of the research area had significant positive global autocorrelation, representing spatial agglomeration characteristics. The spatial characteristics of LISA aggregate map were basically consistent with the intensive degree map; and 3) the labor index, the average mechanical power of the geography and the development rate of agricultural science and technology in some regions, and the effective irrigation rates were positively correlated with the intensive use of farmland. Due to the radiation and dependence of space, the distribution of regression coefficient values has obvious area features of “slices” or “bands”.