A stationary diesel engine connected to an electricity generator was set up to determine the exhaust emissions of polluting gases and opacity when the engine is operated at different engine loads and with different castor-oil-plant biodiesel mixtures, as an alternative to diesel fuel and palm-oil biodiesel blends. The data was employed to model a feed-forward multi-layer neural network, using the software NNModel to train the neural network, and predict the behavior of emissions from the combustion of the fuel based on two input variables: the engine load and the castor-oil plant biodiesel fuel mixture. The results of NNModel were compared against experimental data and analyzed to identify how the engine emissions are affected by the implementation of an alternative biodiesel fuel.