Electrostatic separation is a sustainable dry separation technique based on triboelectric charging of different cellular tissue components (i.e. protein bodies and fibres). This research aimed to determine the mechanism behind the ineffective electrostatic separation of soy and if it could be altered by de-oiling. Several scenarios were compared, which involved electrostatic separation of soy and lupin de-oiled with different solvents (none, acetone, ethanol and hexane). Separation of lupin resulted in a higher true protein content (58.5%DM (dry matter) (Nfactor = 5.7)) than separation of soy (45.0%DM protein). Separation was less effective for soy because its protein bodies were still embedded in the cellular structure after impact milling, which was also reflected in the larger particle size and lower small particle dispersibility. De-oiling soy with hexane and extra milling improved the purity of soy protein-enriched fractions (59.6%DM protein) reaching more similar purity as for lupin protein-enriched fractions. The protein purity of lupin fractions could be increased most via the use of polar solvents (acetone and ethanol). Better electrostatic separation after de-oiling and extra milling for soy could be explained by increased liberated protein bodies, which was key towards the improved separation of soy obtaining protein purities closer to that of pure protein bodies.