Nonaqueous solvent extraction is a promising technology for bitumen recovery from oil sands. In this study, influences of temperature, contact time, stirring rate, and solvent-to-oil sands ratio (V/M) on bitumen recovery, using a mixture solvent of n-heptane and toluene (V/V, 3:1), were investigated by L-9 (3(4)) orthogonal design. Under the orthogonal experiment conditions, the overall impacts of factors were ranked: V/M > stirring rate > contact time > temperature. Profiles of bitumen fractions (saturates, aromatics, resins, and asphaltenes) in the dissolved bitumen, suspended particles, and residual bitumen were investigated in single factor experiments. Asphaltenes have higher temperature sensitivity than other fractions. About 3-7 wt % bitumen particles coexisted with clay minerals (50-70 wt % of the suspended particles) suspended in the solution, most of which were composed of asphaltenes. Approximately 75-90 wt % of SAR fractions (saturates, aromatics, and resins) were dissolved in the composite solvent under the experimental conditions. The amount of residual fractions varied with conditions, and multistage extraction enhanced the bitumen recovery by up to 99%. The evaluation method for bitumen recovery based on dissolved fraction outperformed the methods based on the sum of dissolved and suspended particles. The conceptualization of the solvent extraction process in this study would improve the knowledge base for bitumen recovery mechanism and serve for future work on the engineering applications.