The selection of measures of quality of life used in clinical trials of multiple sclerosis (MS) should be evidence-based. Head-to-head comparison of measures facilitates the selection of measures. The aim of the study was to compare the psychometric properties of the physical and psychological dimensions in three measures of quality of life to aid choice of the most appropriate scale for use in clinical trials of MS. One hundred and twenty-one people with MS (rehabilitation = 57; steroids = 64) completed a selection of health measures before and after treatment. The psychometric properties of three measures of physical function (MSIS-29 physical, SF-36 physical functioning, FAMS mobility) and three measures of psychological function (MSIS-29 psychological, SF-36 mental health, FAMS emotional well-being) were compared by examining data quality, scaling assumptions, acceptability, reliability, validity and responsiveness. Physical (0.63-0.71) and psychological (0.70-0.75) scales were substantially correlated indicating they measure related constructs. The MSIS-29 physical and psychological scales satisfied all criteria for internal consistency reliability (physical = 0.91; psychological = 0.89) and validity. The SF-36 physical scale had a notable floor effect (20%). The FAMS mobility scale had lower reliability (alpha = 0.78) compared to other measures. The MSIS-29 physical (effect size = 0.91) and psychological (effect size = 0.62) scales were the most responsive. In these three samples, the MSIS-29 had better measurement properties for combined physical and psychological health than the SF-36 and the FAMS.