Multi-compartment diffusion imaging using the spherical mean technique (SMT) is a recently developed method that provides more specific information for probing axonal properties in e.g., multiple sclerosis (MS) patients. However, the accuracy and precision of SMT have not been comprehensively investigated to date. We used Cramér-Rao Lower Bound (CRLB) analysis and MS patient data to optimize an SMT acquisition protocol for practical MS imaging. Second, we used computer simulations to comprehensively evaluate the accuracy and precision of SMT for reporting axonal information under various pathologic conditions. The results support better implementation and interpretation of SMT imaging for MS patients.