In unsteady and non-uniform thermal environments, adjusting personal dress states is an effective method to achieve thermal comfort. In this process, local clothing thermal insulation is not evenly distributed throughout the body, which is a key parameter in the human thermoregulation model. However, current thermal comfort standards do not provide a direct way to obtain local thermal insulation values. This study summarizes typical year-round attire combinations and common dress states based on a field study. The local thermal insulation of 40 ensembles and 60 garments of clothing was tested using a thermal manikin. A predictive model for the local thermal insulation of ensembles, based on the overall thermal insulation of garments, was established. The research also found that changes in dress state can significantly alter thermal insulation in areas such as the arms and chest. The JOS-3 and UCB models were used to calculate and found that the lower the ambient temperature, the greater the influence of changes in local clothing thermal insulation on human skin temperature and thermal sensation. A corrected model for different dress states was subsequently developed. Comparison with existing models revealed that the RMSE between the predicted and measured values of the developed model is lower. This study provides more accurate and convenient methods for determining local thermal insulation, offering a data foundation for improving indoor environment control and enhancing energy efficiency.