城市景点是城市对外展示自己的窗口，也是吸引游客和提升城市形象的重要方式之一。营造良好的景点环境，构建满足游客需求的城市景点，是当前景点建设与规划的亟待解决问题。因此分析城市景点的空间分布格局，了解游客对景点的情感倾向并探究不同类型景点影响游客情感倾向的因素，是解决上述问题的关键。本文选取江苏省南京市为例，以网络景点点评数据、城市poi数据、城市景点景观作为研究数据，通过词频分析法提取点评数中游客对南京市景点的感知要素，利用基于机器学习的方法计算出游客的情感倾向，归纳整合多种景点要素数据，构建景点游客情感倾向评价体系，创建南京城市景点游客情感倾向的回归模型。结果表明：南京市景点游客情感倾向总体表现为积极倾向，在空间上分布不均匀且呈现“中心聚集，四周零散”的特征，不同类型的景点游客情感倾向差异较大；游客对景点的服务与管理，景点的环境、景点的历史文化更为关注，对景点的气候关注度相对较少。从总体来说，南京市景点人文景观要素造成景点游客的积极情绪；大多数类型景点步行15分钟范围内生活服务设施越多也让游客产生积极的情感倾向。相比于自然景区类型景点，公园类型的景点游客情感倾向则由多方面造成的，其中植被覆盖情况是造成游客的情感倾向的主要影响因素，并且越靠近市中心景点游客情感倾向越积极；大多数类型景点中，在模型评估方面，在通过参数的微调后，自然景区、公园、现代文化设施、游乐场、休闲度假区等精度分别上升了7.8%、2.30%、8.2%、19.5%、32.2%、27.6%。

City attractions are the Windows for cities to show themselves to the outside world, and also one of the important ways to attract tourists and enhance the image of cities. It is more pressing priority for the construction and planning of scenic spots to create a good environment for scenic spots and to build a city scenic spot that meets the needs of tourists. As a result, the key to track the above problems is to analyze the spatial distribution pattern of urban scenic spots, understand the emotional tendency of tourists to scenic spots, and explore the factors affecting the emotional tendency of tourists in different types of scenic spots. In this paper, Nanjing City of Jiangsu Province was selected as an example, and the online scenic spot comment data, city poi data and urban scenic spot landscape were used as the research data. The word frequency analysis method was used to extract the perceived elements of tourists to Nanjing scenic spots in the number of comments, and the method based on machine learning was used to calculate the emotional tendencies of tourists, and the data of various scenic spots elements were summarized and integrated. To construct the evaluation system of scenic spot tourists' emotional tendency and establish the regression model of scenic spot tourists' emotional tendency in Nanjing. The results show that the emotional tendencies of tourists in Nanjing are generally positive, and the spatial distribution is uneven, It reflects the characteristics of " central aggregation and very little spread around ". The emotional tendencies of tourists in different types of scenic spots are quite different. Tourists pay more attention to the service and management of scenic spots, the environment of city attractions, the history and culture of city attractions, and pay less attention to the climate of scenic spots. In general, the cultural landscape elements of scenic spots in Nanjing lead to the positive emotions of tourists. The more life service facilities within a 15-minute walk, the more positive emotional tendencies of tourists, while the transportation facilities do not necessarily have positive emotional tendencies. Compared with natural scenic spots, the emotional tendency of tourists in park scenic spots is caused by many aspects, among which vegetation cover is the main factor that causes the emotional tendency of tourists, and the closer to the downtown scenic spots, the more positive the emotional tendency of tourists. In most types of scenic spots, the accuracy of model evaluation increased by 7.8%, 2.30%, 8.2%, 19.5%, 32.2% and 27.6%, respectively, for natural scenic spots, parks, modern cultural facilities, amusement parks and leisure resorts after the fine-tuning of parameters.