Abstarct: In the process of electrical machines design, the number of unknown parameters is more than available equations. Therefore, the value of some parameters must be determined by the designer. These parameters are called optional parameters of the machine design. The number of optional parameters in each machine design problem is different. In the design process of conventional machines, the value of feeding voltage and frequency is determined according to the available network. However, the permanent magnet machines are fed by the drive system. Therefore, in the design process of these types of machines, the frequency and voltage are also optional parameters. The unknown values of optional parameters at the beginning of the design process can lead to confuse machine designers. The way of choosing optional parameters has a significant effect on the performance characteristics of the machine. On the other hand, in machine design problems, achieving maximum efficiency, maximum power density, minimum Back EMF harmonics, minimum torque ripple, etc. are usually the design objectives. Therefore, in this thesis, baxsed on analytical methods, a comprehensive design algorithm for the optimal design of a nonslotted axial flux permanent magnet (AFPM) machine is suggested. In addition, some methods for selecting the optimum values of feeding voltage and frequency, reducing the harmonics of Back EMF and the torque ripple of the machine are presented. baxsed on the proposed design algorithm, a sample machine with power of 3.7 kW and speed of 1400 rpm is designed and constructed. With the help of Finite Element (FE) simulation and experimental results the accuracy of the design algorithm and the effectiveness of the proposed methods for optimal selection of feeding voltage and frequency, reduction of Back EMF harmonics and torque ripple of machine are confirmed.