The debate about how best to measure social class in women complicates the analysis of socioeconomic inequalities in women's health. The changing position of women in the labour market may mean that the commonly used "conventional" approach where a woman's partner's occupation is used to estimate her social class may no longer be appropriate. Alternative measures grade a woman's class according to her own occupation or the most dominant class position in the household regardless of gender. We examined the association between "conventional" and personal measures of social class and all-cause mortality in a prospective study of women aged 39-79 years, without prevalent disease, living in the general community in Norfolk, UK, recruited using general practice age-sex registers in 1993-1997 and followed up for an average of 11.9 years. The risk of mortality increased with decreasing social class. There was little difference in the relationship between mortality and social class in women assigned using personal or partner's occupation. When both measures were included in the same model, the effect of both measures was slightly attenuated. We found little difference between the different methods of assigning social class in women and mortality risk prediction. Both measures appear to share some common pathways through which they affect risk of mortality, although confidence intervals were large and neither measure was statistically significant. Further research is needed to disentangle their separate effects and pathways to mortality.