In the context of the artificial intelligence era, machine learning algorithms can be used to achieve rapid prediction of the solar potential of urban blocks. However, most of the existing studies used computer simulation, this paper aims to construct a prediction model of the solar potential of urban blocks based on machine learning algorithms. This paper constructs a parametric model of office blocks based on real block cases in Wuhan city, and develops an arrayed solar potential prediction model for office blocks based on ensemble learning algorithms. The results showed that the ensemble learning prediction model can achieve rapid prediction of solar potential, achieve rapid feedback of solar potential of the design scheme, and significantly improve the efficiency of the performance evaluation and design optimisation of the urban block design scheme; the R2 of the solar potential prediction model for the roof and façade are 0.99 and 0.98, respectively, and the prediction model has a high degree of accuracy. This study aims to provide a decision-making basis for the planning and design of urban office blocks, improve the scientific and rational design, and help urban office blocks achieve carbon neutrality.