Community forestry in the UK has developed rapidly over the last 25 years, and the wide range of drivers has resulted in a great variety of relationships between communities and woodlands, and over 650 community woodlands. Given strong current policy interest, the study aims to assess existing evidence for the impact of these initiatives. The variety of models, evaluation purposes, and impacts requires a new approach to organising the evidence, including a new typology of community woodlands. The review identified more than 70 studies, covering 681 evaluation cases. Of these, 41% are ‘urban regeneration’ programmes, 32% are locally led ‘community place’ projects, and 22% are locally owned ‘community resources’. Only 3% are ‘economic partnerships’ where the primary objective is enterprise; and 1% are ‘lifestyle alternatives’. The majority of evaluations are conducted by the public sector. Evaluations tend to focus on the positive and the quantitative and relate predominantly to outputs (e.g. trees planted, meetings attended). Only 21% of cases identify outcomes (e.g. neighbourhoods enhanced, wellbeing enhanced), and there is little evidence of community empowerment or meaningful engagement in decisionmaking. Attention has shifted from biophysical to social and participation indicators in recent years, but evidence of change over time is lacking. The policy relevance of the evidence base will be greatly enhanced if cases distinguish between types of community woodland, consistently include comparable indicators, and link context, process, outputs and outcomes.