Research on the longitudinal relationship between posttraumatic stress disorder (PTSD) and social support among survivors of large-scale trauma is limited. This study assessed bidirectional relationships between PTSD and perceived social support in a large sample of the 9/11-exposed cohort over a 14-year follow-up. We used data from 23,165 World Trade Center Health Registry (WTCHR) enrollees who were exposed to the 9/11 attacks and participated in the first four WTCHR surveys (Wave 1 (2003–2004) to Wave 4 (2015–2016)). PTSD symptoms were measured using the 17-item PTSD Checklist. Perceived social support was measured using the five-item version of the Modified Social Support Survey. We used a cross-lagged panel analysis and found an inverse relationship between PTSD symptoms and social support. PTSD at Wave 2 (W2) predicted less social support at Wave 3 (W3) (β = −0.10, p < 0.01), and PTSD at W3 predicted less social support at W4 (β = −0.05, p < 0.01). Conversely, social support at W3 buffered PTSD symptoms at W4 (β = −0.03, p < 0.05). Sub-analyses by types of perceived social support suggest greater effects of PTSD on emotional support than tangible support and in community members than rescue/recovery workers. Our findings suggest a bidirectional effect between PTSD symptoms and social support in a longitudinal study of 9/11-exposed populations.