Based on the SPOT-5 images in 2004 and 2010 and SPOT-6 images in 2016 on the northern region of Hexi new town of Nanjing, the urban green space pattern of 2022 was predicted using the CA-Markov model. On this basis, four landscape indexes were used to analyze the evolution of green landscape pattern in the study area from 2004 to 2022. The results showed that the spatial accuracy of urban green space simulated by CA-Markov model was as high as 87.23%, and the Kappa index was 0.8533, which means high reliability. It is predicted that the scale of urban green space in the northern region of Hexi new town will continue to increase in 2016-2022, accounting for 17.28%, and the patch area of the dominant green space will keep increasing. The number of green patches NP and complexity LSI show a downward trend, the green landscape begins to become orderly, the fragmentation is lessening, the green space network system is gradually formed, and the spatial layout of urban green space tends to be reasonable. The simulation results can provide decision-making basis for the dynamic control of urban green space of Hexi new town of Nanjing. Keywords: the spatial accuracy of urban, Kappa index, the number of green patches, landscape index, the remote sensing image data