Background People with severe mental illness often struggle with social relationships, but differences among diagnostic groups are unclear. We assessed and compared objective and subjective social relationship indicators among patients with psychotic, mood and neurotic disorders one year after hospitalisation in five European countries (Belgium, Germany, Italy, Poland and United Kingdom). Methods The number of social contacts, including family members and friends during the previous week (Social Network Schedule), and satisfaction with the number and quality of friendships (Manchester Short Assessment of Quality of Life Quality) were assessed by face-to-face interview. Linear regression models were used to analyse associations with diagnostic groups. Results Participants (n = 2155) reported on average 2.79 +/- 2.37 social contacts overall in the previous week, among whom, a mean of 1.65 +/- 1.83 (59.2 +/- 38.7%) were friends. Satisfaction with friendships was moderate (mean 4.62, SD 1.77). In the univariable model, patients with psychotic disorders reported having less social contact with friends than those with either mood (p < 0.05) or neurotic disorders (p < 0.001), but this difference disappeared when adjusting for socioeconomic and clinical variables (beta = - 0.106, 95% CI - 0.273 to 0.061,p = 0.215). Satisfaction with friendships was similar across diagnostic groups in both univariable (beta = - 0.066, 95% CI - 0.222 to 0.090,p = 0.408) and multivariable models (beta = 0.067, 95% CI - 0.096 to 0.229,p = 0.421). The two indicators showed a weak correlation in the total sample (total social contacts,r(s) = 0.266;p < 0.001; friends,r(s) = 0.326,p < 0.001). Conclusion While objective and subjective social relationship indicators appear to be weakly correlated concepts, there is no variation in either indicator across diagnostic groups when confounders are taken into account among patients with severe mental illness. Interventions specifically targeting social relationships are needed, but they do not necessitate diagnosis-specific adaptations.