A new self-pyrolysis reactor was designed to reduce the utilization of electric heater due to the energy saving for the production of bio-oil from oil palm shell. The yield of the bio- oil was then evaluated with the developed mathematical model by Sharma  with the characteristic of oil palm shell . During the simulation, the temperature on the combustion chamber on the release of the bio-oil was utilized to determine the volatile composition from the combustion of the oil palm shell as fuel. The mass flow was assumed constant for three experiments. The model resulted in a significant difference between the simulated bio-oil and experiments. The bio-oil yields from the simulation were 22.01, 16.36, and 21.89 % (d.b.) meanwhile the experimental yields were 10.23, 9.82, and 8.41% (d.b.). The char yield varied from 30.7 % (d.b.) from the simulation to 40.9 % (d.b.) from the experiment. This phenomenon was due to the development of process temperature over time which was not considered as one of the influential factors in producing volatile matters on the simulation model. Meanwhile the real experiments highly relied on the process conditions (reactor type, temperature over time, gas flow). There was also possibilities of the occurrence of the gasification inside the reactor which caused the liquid yield was not as high as simulated. Further simulation model research on producing the bio-oil yield will be needed to predict the optimum condition and temperature development on the newly self-pyrolysis reactor.