Slotting directional hydraulic fracturing is a new method for improving permeability of a coal seam in underground coal mines that can solve problems related to non-uniform permeability enhancement in the seam. A slotting nozzle is the key to this technology: its performance determines the length and stability of the slotted hole. In this study, computational fluid dynamics was used to study the effects of stable segment length L, convergent angle θ, and straight segment length l on the performance of slotting nozzles. The results showed that the order of parameters affecting nozzle performance is: θ > L > l. When the total length of the slotting nozzle was fixed, the dynamic pressure gradually decreased with an increase of L and the rate of decrease slowed down. With an increase of θ, dynamic pressure increased quadratically and the rate of change gradually decreased. With an increase of l, the dynamic pressure decreased quadratically and the rate of change gradually increased. A slotting nozzle (L = 0 mm, θ = 30°, l = 9 mm) was manufactured and measurements of flow coefficient, water jet morphology, and a slotting experiment were carried out. The experimental results showed that the flow coefficient, water jet convergence, and slotting depth of the optimized slotting nozzle were obviously higher than those of the original design, which proved the validity of using this numerical method to optimize the slotting nozzle structure.