In the background of global warming, terrestrial ecosystem carbon cycling plays an important role in the change of atmospheric CO_2 concentration. The current large areas of forest and forestry or agroforestry is pivotal to the balance between economic development and ecological maintenance. From January to December 2014, we measured using an automated soil CO_2 flux system (LI-8100, LI-COR, USA) the soil respiration of four models (walnut artificial pure forest (CK), walnut - Radix Chuan Mingshinis composite model (HC), walnut - Platycodon compound model (HJ), and walnut - Agastache rugosa composite model (HH)). The soil temperature and humidity were measured at the same time. The measurements showed a fi rst increasing and then decreasing monthly variation trend in soil respiration. The soil respiration was the maximum in late June for HC, but in mid-July for CK, HJ and HH. The lowest soil respiration of HJ, HH, HC were 1.36, 1.61 and 0.89 times r espectively higher than that of CK, with the lowest value in January. There was a signifi cant exponential relationship between soil respiration and soil temperature, and a reciprocal relationship between soil moisture and soil respiration, but the relationship was signifi cant only in HH, indicating that HH models should be more conducive than others to reduce soil carbon emissions. The research showed that soil temperature may be the main factor of soil respiration in the hilly region of Sichuan Basin, and the operating model may be an important infl uencing factor for the role of carbon source/sink in the study area.