The stability of lipid nanoparticles (LN) was evaluated using a Central Composite Rotatable Design (CCRD) full factorial type 22 composed of four factorial, four axial, and three central points, totaling 11 experiments. The independent variables were concentration of whey protein used as an emulsifier (0.5 to 3.0%) and concentration of triolein added to the lipid phase (0.0 to 95.0%). The LN were obtained from oil-in-water type pre-emulsions (10% lipid phase), which were homogenized under high pressure (NS 1001 L - Panda 2K, 70 ºC, 800 bar and 3 homogenization cycles). The particle size and polydispersity index were used to evaluate the stability of the system, using Mastersizer 2000 Laser Diffractometer and creaming index for 7 days. The models were evaluated in relation to the coefficient of determination and F-test at 5% of significance. The particles had an average diameter of 155 ± 16 nm, with no significant differences in size for all conditions evaluated, indicating that the homogenization condition was adequate for obtaining LN, which should present a diameter < 200 nm. Both the concentration of whey protein and triolein added to the lipid phase significantly affected the polydispersity index. Lower polydispersity was obtained with whey protein concentrations ranging from 1.7-2.3% and up to 30% triolein in the lipid phase. For the same concentration of proteins (0.87 or 2.60%), LN with concentrations higher than 30% in the lipid phase showed higher polydispersity index, probably due to the increase in low melting point triacylglycerols in the blend. Overall, LN showed high stability to creaming for 7 days. Lower instabilities (maximum 5%) are related to high concentrations of triolein in the lipid phase (95%) or low protein concentrations (0.5%). The results demonstrated that model lipid systems with up to 30% triolein can be stabilized with approximately 2% whey proteins.