Potential uses for electromagnetic launchers in defense systems, space exploration, and transportation have recently emerged. In addition, this accelerator has many applications, such as deploying small satellites into low-earth orbit and accelerating high-speed trains (e.g., bullet trains and Hyperloop) with a low-cost propulsion system instead of expensive linear motors, particularly in space applications. Therefore, the full capability and optimization of these launchers’ efficiency are still required. Therefore, this paper focuses on presenting a new design to decrease the coil’s magnetic circuit reluctance and boost the magnetic flux lines by adding a laminated iron yoke surrounding the coil. This design makes the inductance value of the iron-yoked accelerator twice the inductance in case of the absence of the iron-yoke at its peak. Additionally, the initial inductance of the iron-yoked accelerator is approximately 65% higher than that of the coil without the iron yoke. Consequently, the modified design proposed an efficiency of 17.5%, which represents a 60% improvement over the efficiency of the regular accelerator. In addition, the introduced design eliminates the suck-back force using a fast-switching device (IGBT) to switch the coil off when the projectile reaches half of the coil. Moreover, a mathematical model for the iron-yoked accelerator is built on MATLAB Simulink and validated experimentally. An artificial intelligence optimization technique, the gravitational search algorithm (GSA), is used to optimize the accelerator parameters, such as the number of turns, capacitor value, and capacitor voltage. Finally, the experimental evaluation of the GSA-optimized system demonstrated an additional 15% enhancement in efficiency, bringing the total efficiency to 20%.