Evaluation of bitumen content in oil sands feedstock is important to control the extraction process and improve the recovery efficiency. Current standard techniques for bitumen determination in oil sands ore suffer whether from their time-consuming labor intensive method of sample preparations or from requiring a high number of samples for calibration purposes. The analysis of bitumen ore samples is very challenging for the analytical chemist due to its nature of several phases of wet and dry particles and heterogeneous mixture of clay, bitumen, water, and solid contents. In this paper, we present a new enabling method using LIBS to determine rapidly and without sample preparation the bitumen content in oil sands ores. A qualitative study by principal component analysis was first done, and then a partial least-squares method was performed to assess the feasibility of determining bitumen by LIBS. The results show a good correlation between LIBS spectra and bitumen content and a prediction averaged absolute error around 0.7%. This method demonstrates that the LIBS is a promising tool for rapidly assessing oil sands ore grades either in the lab, at-line, or online. Elemental composition of solids is also investigated.