Abstract: Abstract: Heat transfer coefficient is one of the most crucial parameters in thermal calculation and design for an externally heated rotary kiln. Suitably designed kiln dimensions, structure and operating parameters rely on the accuracy of the employed heat transfer coefficient. For an externally heated kiln, heat transfers from an outside source to inside particles through a wall. Generally, the filling ratio in an externally heated rotary kiln is low. So, the heat transfer mechanism for large particles with a low filling ratio in an externally heated rotary kiln is quite different from that in an internally heated rotary kiln, whose filling ratio is usually more than 15 percent. Despite the existence of some achievements in particles motion behavior and heat transfer mechanisms in an internally heated rotary kiln, so far, there is no reliable heat transfer model to describe the heat transfer process between the kiln's surface and particles in an externally heated rotary kiln with low filling large particles. As a result, the main approach of heat transfer coefficient determination is still an experimental test. On the basis of heat transfer mechanism analysis, this paper regards the heat transfer process between the kiln's surface and large particles as consisting of heat conduction between the kiln's surface and gas film, heat convection between the gas film and particles, and heat radiation between the kiln's surface and particles. Finally, a mathematical model is created for the prediction of the heat transfer coefficient between the kiln's surface and large particles. To validate the developed model, a series of experimental tests are performed. Alumina spherical grains with a diameter of 6 mm are used as testing particles. When the filling ratio is 5 percent, the heat transfer coefficients are measured in the range of 220℃-420℃ at 20℃ surface temperature intervals, corresponding to the rotary speeds of 1r/min, 2r/min, and 3r/min, respectively. The tests find that the heat transfer coefficient only slightly increases with rotary speed increase. However, the coefficient increases intensely when the kiln's surface temperature increases. Comparisons of the experimental results and predictions show that the maximum relative error (emax) is about 9.8 percent, and the average error (eave) is 5.86 percent. According to the engineering design heat transfer coefficient model experience, the model is able to well match the engineering requirement that it can refer to thermal calculation. The results also show that, for the testing material in this paper, the fraction of radiation heat transferred from kiln's surface to particles is more than 75 percent of the total heat transfer when the surface temperature is higher than 220℃. If the surface temperature is beyond 320℃, a more intense increasing percentage of cure will appear. The error analysis shows that the prediction values are all larger than testing results, which could be caused by the assumptions of both particle distribution and radiation heat transfer between the kiln's surface and particles. To obtain a more accurate heat transfer coefficient model for large particles with low filling ratio in an externally heated rotary kiln, it is necessary to carry out further investigate into the performance of the motion behavior of particles. The achievement in this paper is helpful for further investigation of heat transfer mechanism in an externally heated rotary kiln.