This work reports phase equilibrium experimental data (cloud points) for the system annatto seed oil in compressed carbon dioxide + ethanol as co-solvent. The oil was obtained by supercritical fluid extraction at 40 degrees C and 20 MPa using carbon dioxide as solvent. The synthetic static method was employed for the measurements, which were carried out in a temperature range from 25 to 60 degrees C and up to 19.32 MPa. The experiments were carried out using (CO2 + ethanol) overall mass fraction compositions ranging from 0.82 to 0.99. Experimental results were modelled using artificial neural network (ANN). Results indicate a complex phase behavior for all temperatures investigated with the occurrence of vapor-liquid, liquid-liquid and vapor-liquid-liquid transitions. ANN modelling was an efficient alternative for the thermodynamic models to accurately analysis of the phase equilibrium at high pressure.