With the acceleration of global urbanization, the construction of low-carbon cities faces many urgent problems such as resource waste, environmental pollution and traffic congestion. This paper aims to construct a multi-objective decision-making model for low-carbon urban renewal paths based on machine learning algorithms to optimize resource allocation and enhance urban sustainable development capabilities. This paper determines the key factors affecting low-carbon urban renewal through a literature review; secondly, cluster analysis is used to classify urban areas to identify the needs of different characteristic areas; then a multi-objective optimization model is constructed, and the objective function is solved in combination with a genetic algorithm. The optimization objectives include minimizing carbon emissions, maximizing economic benefits and improving social benefits. Through actual case analysis, the application of the model in a certain city shows that the average carbon emissions of the low-carbon urban renewal path are 140.5 tons, and the economic benefit increase reaches 16.64%. The multi-objective decision-making model based on machine learning provides effective path selection and decision support for low-carbon urban renewal, which has important practical significance.