In order to more effectively utilize passive technology to improve the energy efficiency of near zero energy buildings and improve the performance design scheme of near zero energy buildings in China. This study takes an office building in Taiyuan City as a case study, establishes a near zero energy consumption building model, and uses sensitivity analysis method based on Monte Carlo simulation to conduct in-depth research on the parameters that affect building energy consumption; Subsequently, a neural network model was used to establish a prediction model for input parameters and building energy consumption. Genetic algorithms were used to optimize building parameters and solve the optimal combination of near zero energy building design parameters, providing important theoretical guidance for future near zero energy building design.