Background: Cross-sectional studies suggest that more green space may lower the odds of prevalent diabetes, hypertension and cardiovascular diseases (CVD) in cities. We assess if these results are replicable for tree canopy exposure and then extend the study longitudinally to examine incident cardiometabolic outcomes. Methods: The study was set in the Australian cities of Sydney, Wollongong and Newcastle. Total green space and tree canopy as percentages of landcover within 1.6km (1 mile) from home were linked to a residentially stable sample of 46 786 participants in the Sax Institute's 45 and Up Study (baseline 2006-09; follow-up 2012-15). Separate multilevel models were used to investigate whether the odds of prevalent and incident doctor-diagnosed diabetes, hypertension and CVD were associated with total green space and tree canopy provision, adjusting for age, sex, income, education, employment and couple status. Results: Lower odds of prevalent diabetes were observed with 1% increases in total green space [odds ratio (OR) 0.993, 95% confidence interval (CI) 0.988 to 0.998] and tree canopy (0.984, 0.978 to 0.989). Lower odds of prevalent CVD were found with a 1% increase in tree canopy only (0.996, 0.993 to 0.999). Lower odds of incident diabetes (0.988, 0.981 to 0.994), hypertension (0.993, 0.989 to 0.997) and CVD (0.993, 0.988 to 0.998) were associated with a 1% increase in tree canopy, but not total green space. At ≥30% compared with 0-9% tree canopy, there were lower odds of incident diabetes (0.687, 0.547 to 0.855), hypertension (0.828, 0.719 to 0.952) and CVD (0.782, 0.652 to 0.935). However, ≥30% compared with 0-4% total green space was associated with lower odds of prevalent diabetes only (0.695, 0.512 to 0.962). Conclusions: Restoring local tree canopy in neighbourhoods may help to prevent the incidence of cardiometabolic diseases.