Stable water-in-oil emulsions persist in bitumen froth from surface mining process of Athabasca oil sands because of asphaltenes and clay solids. This dissertation focuses on the characterization and separation of water in diluted bitumen emulsions. A novel approach to process experimental data from classic NMR experiments for the characterization of water in diluted bitumen emulsions has been proposed and tested. NMR PGSE restricted diffusion measurement can characterize emulsion drop size distribution. Experiments show that drop size of emulsion does not change much with time, which indicates that water in diluted bitumen emulsion is very stable without demulsifier. Water fraction profile and water droplet sedimentation velocity can be obtained from MRI 1-D T1 weighted profile measurement. Emulsion flocculation can be deduced by comparing the sedimentation velocity from experiment data and Stokes Law prediction. PR5 (a polyoxyethylene (EO)/polyoxypropylene (PO) alkylphenol formaldehyde resin) is an appropriate demulsifier for water in diluted bitumen emulsion. Almost complete separation can be obtained in the absence of clay solids. For the sample with solids, a rag layer containing solids with moderate density forms between the clean oil and free water layers. Partially oil-wet clay solids prevent complete separation of the emulsion. Experiments reveal that wettability of clay solids has significant effect on emulsion stability. Kaolinite with 100 ppm sodium naphthenate in toluene-brine mixture is chosen as model system for wettability test. Wettability of kaolinite can be altered by pH control, silicate and surfactant. Adding 3x10 -3 M Na2SiO3 at pH 10 can get 80% of kaolinite water-wet. Over 90% of kaolinite becomes water-wet adding C8TAB, betaine 13 and amine oxide DO with optimal dosages. In diluted bitumen emulsion, about 10-4 M sodium meta-silicate can change the wettability of solids from partially oil-wet to more water-wet. Hereby the clay solids can settle down to the aqueous phase and the separation is almost complete. Wettability of kaolinite can be characterized via zeta potential measurement and modeling. Simplified Gouy-Stern-Grahame model and oxide site-binding model can be used to correlate zeta potential of kaolinite in brine with different additives. Sodium silicates have the greatest effect per unit addition on changing zeta potential of kaolinite and can be used to change the wettability of clay solids. Almost complete separation be obtained by the three-step procedure: (a) adding 10-4 M Na2SiO3 during initial emulsion formation to make the solids less oil wet; (b) removing the clean oil formed following subsequent treatment with demulsifier and adding NaOH or Na2SiO3 with shaking to destroy the rag layer and form a relatively concentrated oil-in-water emulsion nearly free of solids; and (c) adding hydrochloric acid to break the oil-in-water emulsion.