Background: Studies investigating the potential role of neighbourhood green space quality on the development of prosocial behaviour among children are sparse. This study aimed to investigate the longitudinal association between caregiver perceived green space quality and child prosocial behaviour, and identify potential effect modifiers of the association. Methods: This was a longitudinal study using data from the Longitudinal Study of Australian Children, involving 4,969 children aged 4–5 years that were biannually followed-up from 2004 to 2014. Prosocial behaviour was assessed using a prosocial scale from Goodman's Strengths and Difficulties Questionnaire. Green space quality was measured based on caregiver perception of the availability of neighbourhood parks, playgrounds, and play spaces of good quality. Latent class analysis was used to partition children into groups denoting different levels of caregiver perceptions of green space quality accumulated over 10 years. Multinomial logistic regression was used to examine the likelihood of being in groups with favourable perception of green space quality. Multilevel linear regression was used to examine associations between trajectory groups and prosocial behaviour. Separate multivariate models were developed to assess the potential role of quality green space in reducing prosocial behaviour related inequalities. Furthermore, two-way interaction terms were added into the models to identify potential effect moderation. Results: There were six trajectory classes of green space quality perceived by caregivers. The likelihood of being in groups with better green space quality varied by neighbourhood circumstances. Children with consistently very good quality green space had higher prosocial behaviour (β = 0.35; 95%CI = 0.23, 0.47) than those with low quality green space. Better prosocial behaviour was also observed among children whose caregiver perception of green space quality trended from good to very good (β = 0.23; 95%CI = 0.11, 0.35) and from very good to good (β = 0.31; 95%CI = 0.20, 0.42) compared to children with consistently low quality green space. Very good quality green space perceived by caregivers over time potentially reduces socioeconomic inequalities in prosocial behaviour. Green space quality-prosocial behaviour association was stronger among boys, children speaking only English at home, those living in more affluent areas, and remote areas. Conclusion: Trajectory of caregiver perceived green space quality was positively associated with child prosocial behaviour. The findings suggest that improving the quality of green space to be very good quality, particularly in deprived and less accessible areas may help improve prosocial behaviour in children.