Study on ecological networks about species interactions provides a basis for species diversity protection, and it is beneficial for the maintenance of urban ecosystem stability. However, urban ecological network study between butterfly and plant based on community level still leaves a huge blank, therefore, the knowledge of butterfly-nectar plant network structure needs an in-depth understanding. To our knowledge, ours is the first study focusing on butterfly-nectar plant network in urban ecosystems in China and evaluating the effects of different type of plants on butterfly-nectar plant network. In this study, we surveyed flower visiting butterfly species and nectariferous plant species and the interaction frequency in 26 urban parks in Beijing, China from June to September in 2020. The butterfly-nectar plant bipartite network was built. We identified five quantitative metrics on two hierarchical levels (network and species) to describe network patterns and species' interaction. Metrics on network level were interaction diversity (ID), interaction evenness (IE) and network specialization (H2'). Moreover, we analyzed the effect of different vegetation types on the network structure differences, including different type of growth form, origin, and cultivation plant, using Kruskal-Wallis test and variance partitioning analysis. Metrics on species level were partner diversity (PD) and species specialization (d') for identifying important nectariferous plant species that provide a key and central role to the stability of the structure of network. The results are as follows: (1) 22 species of butterfly and 81 species of flowering plants of interaction were observed and the butterfly-nectar plant network tended to be generalized. (2) The interaction diversity and specialization of two growth form plant-butterfly network were significantly different (P<0.01), and same to of two origin plant. Furthermore, herbs and native plants played an important role in maintaining the interaction diversity and specialized butterfly species in the network. Interestingly, cultivation type of plants had little effect on nectar network structure. (3) Plants with high partner diversity and high degree of specialization could be regarded as important nectariferous plant species. Consequently, in order to protect butterfly diversity, native herbaceous plant community should be constructed in green space of urban ecosystem, and important nectariferous plant species should be given top priority to grow. Our findings confirm that the butterfly-nectar plant ecological network can provide a scientific strategy for biodiversity conservation in urban ecosystems. It could be an effective tool to link ecology research with urban green space practice management.