In order to reduce the excessive nitrogen oxide (NOx) emissions from traditional swirl burners in heat treatment furnaces, a low nitrogen burner optimization method based on gas classification technology was proposed. Firstly, the ordinary swirl burner was used as the basic structure, and the gas grading structure was added without significantly increasing fuel consumption. Secondly, the ANSYS Fluent software system was used to simulate the influence of the blade angle and air flow rate of the first-stage stator cascade on the NOx generation. Finally, by comparing the mixing uniformity, temperature distribution and NOx concentration field, the optimal parameter combination was determined. The results show that the gas grading structure significantly improves the mixing efficiency of secondary gas and swirling air, the heat load distribution is more uniform, and the thermal NOx production is reduced. When the first-stage stator cascade blade angle is 30°, the air flow rate is 1.15 m/s, and the proportion of primary gas is 65%, the burner performance is optimal, and the NOx emission is significantly reduced compared to the traditional structure. This study provides an optimization direction for the design of low nitrogen burners and has reference value for the environmental protection transformation of industrial combustion equipment.