Urban human settlement is a critical criterion in judging the developmental level of a city. It is safe to say that it is the reflection of urban economic circulation and environmental construction in the sense of geographical space under urbanization. It is an important subject to evaluate the degree of urban human settlement, to analyse the problems in various cities, to make a classification in the angle of self-realization, and to explore its regional similarities, characteristics and reasons that cause the distinction. The self-organization feature maps network model was firstly applied to the evaluation of main urban human settlements in China. Urban human settlement is not only a critical criterion in judging the level of development of a city, but also a possible tool to make a classified judgment on urban human settlements in 35 of main cities from different angles such as economic realization, digital realization, acquisitive realization, residential realization, transporting realization, security realization, virescence realization, recreational realization and natural environmental realization, to distinguish the differences in different cities, so as to provide basis for construction and management of urban human settlements in different cities. Based on the research, the urban human settlements in China is provided with the particularity of obvious regional distinctions: its quality in littoral cities is mostly higher than that in the middle and western regions. It shows that the geographical elements count fundamentally in urban human settlements, and economic capacity is highly connected with it. The top seven cities in the evaluation of urban human settlements all come from the developed eastern littoral, for the prosperity of the cities provides more job opportunities and better surroundings. Meanwhile, the funds allocated to the construction of the infrastructure and the melioration of environmental conditions need the support of prosperous economy. From the view of classified evaluation, the self-realization of people requires more on economic realization, digital realization, residential realization, transporting realization, security realization, virescence realization, and recreational realization. In other words, in urban human settlements, the economic environment, salary, circulation of information, degree of modernization, virescence condition and the comfort of living are the main ingredients when choosing urban human settlements in the process of self-realization. This model based on the SOM neural network is applied to quality evaluation of main urban human settlements in China. The evaluation value indicates the current situation of urban human settlements and relative levels in China.The run results of the model show that the application on quality evaluation of urban human settlements based on the SOM neural network is convenient, precise and feasible which proves to be a valuable evaluation approach.