The causes of frequent haze in Beijing-Tianjin-Hebei were truly complex, but the researches on the interrelation between haze and socio-economic factors were still lack of quantitative analysis. Considering that the interrelation under different regimes showed different change rules, the modified panel threshold regression model was applied to address different policy challenges. The results showed that the single threshold effect was significant while the added value of the tertiary industry or public transport was selected as the threshold variable. When the added value of the tertiary industry was less than the threshold value, the development of the secondary industry and the rising number of cars would increase PM_2.5. Therefore, the expansion of the public transport would be an effective measure to ease the haze. When the added value of the tertiary industry exceeded the threshold value, the continued development of the secondary and the tertiary industry would reduce PM_2.5, but the relief function achieved by expanding the public transport would be weakened slightly. Meanwhile the number of vehicles would become the primary factor influencing PM_2.5. When the number of public transport was closer to its threshold value, the opportunity cost of PM_2.5 reduction would be less and the increase of energy consumption would always increase PM_2.5. As there were nonlinear relations between the haze and the socio-economic factors, the policy formulation of PM_2.5-reduction should be combined with the threshold effect.