A pilot study of the effects of metacognition-oriented social skills training (MOSST) on social functioning in patients with schizophrenia spectrum disorders (SSDs) reported promising results. The main purpose of the current trial was to compare the effectiveness and potential benefits of MOSST vs conventional social skills training (SST). Single-blind randomized controlled trial with 2 groups of patients aged 18–65 with SSDs on partial hospitalization. Participants were randomly assigned (1:1) to receive 16 group sessions with MOSST or conventional SST, both in addition to standard care, over 4 months, with a 6-month follow-up. Psychosocial functioning, metacognition, and symptom outcomes were measured by blind assessors. Statistical analyses used mixed models to estimate treatment effects in each postrandomization time point. Thirty-six patients were randomly assigned to the MOSST group and 33 patients to the conventional SST group. Between-group differences were significant in favor of MOSST on Social and Occupational Functioning Assessment Scale (SOFAS) and Personal and Social Performance Scale (PSP) total scores at post-treatment and follow-up. Concerning PSP subscales, there were significant between-group differences in favor of MOSST at follow-up on socially useful activities, personal and social relationships, and disturbing and aggressive behaviors. Metacognition only improved following MOSST group. For people with SDDs, MOSST appears to have short- and long-term beneficial effects on social functioning and symptoms. Further studies are required to replicate the current results in other samples.