The urban world of the 21st century is composed of numerous nodes, streams and webs, which create a new landscape of globalization and impose different logic of space and time perception. Therefore, the urban infrastructure is updated and its networks are continuously multiplied. A method known as urban acupuncture on the one hand tests the local effects of every project, and on the other hand establishes a network of points or dots to act upon. The main objective of this paper is to relate the concept of urban acupuncture with the use of a neural network algorithm to determine those points where developing actions in order to improve the quality life in cities. We apply the neural network model GNG3D to the design of a simplified network in a real city of our surrounding.