The study on the structure of complex space ecological network with spatial attribute information is a blank of the current research on complex network. Dengkou County, a typical county in the desert oasis area, was selected as the study area. The method of point pattern analysis in complex network analysis methods, the algorithm of extracting network skeletons and the robustness index of spatial structure were modified in accordance with the characteristics of the actual spatial ecological network, to analyze the structure of spatial complex ecological network in the study area. The results showed that the nodes of seven types of eco-sources exhibited aggregated spatial distribution pattern at a small scale, which was gradually distributed randomly as the scale increase, and finally distributed uniformly. This distribution feature ensured that ecological networks were stable at both small and large scales. The NDVI and MNDWI values of different types of ecological source had different distribution characteristics. The density of water corridors was ranged from 0 to 1.46, and the density of road corridors was ranged from 0 to 2.27. The framework corridor identified by Kruskal algorithm combined with the ecological network features was in accordance with the actual situation. The initial robustness of the current ecological network was only 0.73, and the robustness of the ecological network after optimization reached 1. Malicious attacks were more destructive than random attacks, but the optimized ecological networks showed stronger antistrike capability, stronger connectivity and stronger robustness of node and edge recovery. After optimization of the ecological node layout, the ecological network nodes and edges were more resistant to attack and destruction and more resilient.