Although heavy oil is an abundant and promising energy source, its processing and utilization are complicated due to its high density, low hydrogen/carbon ratio, and high asphaltene content. Fortunately, these problems can be mitigated by the application of oxy-fuel combustion. To gain deeper insights into the above technology, the characteristics of heavy oil combustion in an O-2/CO2 atmosphere was investigated using non-isothermal thermogravimetric analysis. We demonstrate that the combustion process consisted of four stages. Low-molecular-weight hydrocarbons reacted at low temperature, whereas heavy ones required a higher temperature. Increasing the concentration of oxygen resulted in increased TGA and DSC peak intensities and decreased peak widths, and these peaks were shifted to lower temperatures. Coat-Redfern and Flynn-Wall-Ozzawa methods were used to evaluate the kinetic parameters (E, A) of the oxidation process, showing that the high-temperature activation energy was much higher than the low-temperature one due to the different molecular weights of the oxidized substrates in each region. The reaction was demonstrated to be diffusion-controlled, as reflected by the lower activation energy at high oxygen concentration and high temperature, with the influence of oxygen concentration on QO processes being much more obvious than that on SO ones.