Rapid urbanization has being highlighted the limitations on the sustainable development of cities due to the fragmentation of restricted urban green space in recent two decades. Green infrastructure (GI) is considered generally to involve all natural, semi-natural, and artificial ecosystems within and between urban areas at all spatial scales. The concept of GI emphasizes the quality and quantity of urban and peri-urban green spaces, as well as their multiple functions and the importance of interconnections between them. GI does not mean to build an entirely new landscape system, but enhance the connectivity of already existing green spaces and build strong ecological services of ecosystem. In the case of strong limited urban space, it is extremely important to build effectively urban green infrastructure networks based on evaluating its spatial priority. Constructing green infrastructure requires consideration of effectiveness and efficiency of the GI network, as well as current situations of the city to ensure operability. MSPA is a morphological spatial pattern analysis approach that divides the green space into seven categories that do not overlap each other through a series of image operations, which can quickly identify the important structural elements relating to the GI network. "Space syntax" proposes a series of morphological variables, such as "connection value", "depth value", "integration degree", "comprehension value", which are quantitative descriptions of the spatial structure. The combination of them can provide a new perspective for the spatial configuration quantization and spatial priority recognition of GI. In order to provide new modeling and analysis concepts for the construction and management of urban green infrastructure network, this paper takes Pukou District of Nanjing being rapid urbanization now as an example. Firstly, selecting the habitat patch with the greatest contribution to maintain landscape connectivity as hubs of the green infrastructure network through applying the MSPA method combined with the landscape connectivity index. Then constructing the potential GI corridors through applying the minimum green path method. Thirdly, identifying the priority of "Green Trail" from the perspective of spatial accessibility using the space syntax, and optimizing and controlling GI network planning combined with the urban ecological red line and green space system planning. Finally proposing optimization approach to green network for spatial structures that make the construction of GI network more scientific and easier to implement in a focused and staged path. The results show that:(1) Pukou District is the core area of Nanjing Jiangbei New District (National New District), Where has a big gap for urban construction land use and it is not practical to carry out completely the urban ecological construction on large scale. Therefore, it is economically feasible to use the existing resources to optimize the urban green infrastructure network by extracting important core areas and bridge areas through the MSPA method. (2) The syntax variables, such as "general integration value" and "choice", are used as quantitative values to evaluate the priority of GI spatial structure based on the theory of space syntax. These variables provided an important reference for the network optimization of GI. The method combining MSPA, landscape connectivity and space syntax quantifies the spatial characteristics and provides a new research path for network optimization of urban green infrastructure. The research would do good to understand the spatial distribution and priority assessment for green infrastructure network planning, and provide a model for other cities in the course of rapid urbanization to build ecological security patterns and green space systems in urban areas.