The transformation of old neighborhoods is related to the quality of life and happiness of neighborhood residents, and the needs of residents must be fully considered in the development of transformation strategies. The problem to be transformed according to the needs is summarized in this paper, and the key indicators of the neighborhood’s spatial pattern are abstracted. Under the guidance of the perception-behavior correlation model, the correlation and regression curves between spatial indicators and behavioral density are estimated, and the optimal value of the correlated behavioral density is calculated to provide data support for intelligent transformation strategies. The CRITIC method and the cloud model are combined in the article to create an evaluation model for the transformation effect. Taking Kangfu Village in Nanjing as an example, the perception-behavior correlation analysis points out that each public space perception index has an extremely significant statistical correlation with residents’ spontaneous activity density and long-time activity density. The best-fit function curves were obtained through further curve estimation regression analysis, and the ideal value intervals were determined for each perception indicator. After the implementation of the retrofit strategy, the expectation value of the owners’ cloud evaluation of the “community intelligent retrofit” is 3.677, and the similarity between the comprehensive evaluation of satisfaction and the comparative satisfaction level reaches 0.905, which makes the evaluation result more satisfactory. There are sufficient reasons to believe that the intelligent transformation strategy proposed in this paper is generally accepted, and the formulation of the strategy based on the perception-behavior correlation is scientific.