In-cylinder flow characteristics, which are mainly determined by the intake ports, have a decisive influence on the mixture formation and combustion process of internal combustion (IC) engines. To develop the optimum intake ports in a four-valve diesel engine, the full-parametric design method coupled with genetic algorithm (GA) and artificial neural network (ANN) was adopted in this study. Different from previous studies, the full-parametric templates for the tangential and helical ports were established. Then, to improve the precision of the CFD result which was used to train the ANN, the near-wall region affected by viscosity was resolved by the “enhanced wall treatment” during the simulation process. Besides, considering the interaction between the outlets of adjacent intake ports in the four-valve engine, the ANN was divided into three parts to better predict the flow performance of the combined intake ports. Finally, the developed intake ports based on ANN and GA were machined to the flow box and validated by the steady flow test. The low deviation of 4.71% in swirl ratio between the design target and the experiment result confirms that this method can be well applied to the development of the intake ports in a four-valve diesel engine.