The integration and optimization of natural protected areas (IONPA) is a significant way to reconstruct the system of natural protected areas, connect the delimitation of three control lines, and improve the project of land spatial plans in China. In this paper, while using the result of IONPA and the data of the third national land survey, we analyzed the pattern of variations and rationalities of the landscape in each stage following the practice of IONPA in Sinan County, Guizhou Province, with the methods of GIS spatial distribution, landscape pattern index, and boundary analysis. The results showed that after IONPA, (1) the composition of land use/cover varied significantly, the area of cultivated land and construction land decreased dramatically, and the area of forest land from the dominant landscape increased relatively. (2) The landscape heterogeneity decreased, but both the dominance and the fragmentation increased in the natural protected areas in Sinan County. (3) The complexity and heterogeneity of each patch type decreased, but the dominance of forest land increased, while the fragmentation for the cultivated land and garden land increased. The IONPA did not fragment the dominant landscape. (4) The boundary complexity of the natural protected areas increased significantly, where the nature reserve and geopark varied the most, which would increase the difficulties of delimiting and staking the boundaries. (5) The changes of the cultivated land, construction land and forest land were the main factors attributed to the variations of landscape pattern in Sinan County. The cultivated land and construction land were suggested to move out from the natural protected areas mainly due to conservation conflicts, while the forest land was also suggested to move out mainly as fragmentized patches along with conflicts. In this study, the scheme of IONPA was adjusted focused on resolving the contradictions and problems, therefore, the changes of landscape pattern reflected from the practice in this study were consistent with the specific contents of the adjustment rules of IONPA. In general, the scheme of IONPA in this study was feasible and reasonable. However, there are still some problems like fragmentation, boundary complexity and regional separation, which remind us to avoid the above-mentioned problems, and meanwhile, to further optimize the adjustment rules of IONPA and to unify the accuracies of all data in the future. The information coming from the changes of landscape pattern indexes can provide a certain reference to conduct IONPA scientifically and to assess the result rationally. Additionally, introduction of the achievement evaluation mechanism to IONPA was significantly important.