The productivity of agroforestry systems (AFSs) and the provisioning of associated ecosystem services (ESs) are threatened by increasing cropping intensification and climate change. Compared to full-sun coffee, maintaining shade cover might protect against climate variability, forest degradation, and pests/diseases attack. However, there may be trade-offs between yields and ESs. While the impacts of shade have been reviewed, a global synthesis to understand how shade trees and a range of practices and biophysical factors contribute to yield gaps and ESs across management types is lacking. This research aims at integrating quantitative data on management practices (shade cover percentage, tree/coffee plant densities, and nutrient inputs), coffee/tree characteristics, soil properties, topographic attributes and climate, in order to compare the responses of coffee yields, biodiversity and ES supply across management types. We conducted a meta-analysis with a total of 142 papers that fulfilled the inclusion criteria and analyzed the individual and combined effects of these factors. The results show that factors such as shade cover and tree/plant densities affected more the yields than climatic and biotic factors. AFSs under moderate shade cover levels (35–50%), with tree densities between 100 and 250 trees ha−1, at low altitudes, and with steeper slopes provided as much yields as full-sun systems (4.1 ± 2.88 tons ha−1 yr−1). While increased shade cover (>51%) decreased yields, the results also show that more diversified AFSs can support biodiversity conservation and provide ESs such as stored carbon, infiltration, pollination, and water runoff. The generalized linear models (GLMs) analyses showed that coffee yield variations are highly context-dependent and factors such as the incidence of pest/leaf rust, altitude, soil quality, available water, and biodiversity may play an important role. The novelty of this research is that the effect of shade, coffee plants, and input management practices is taken into account, along with the effects of a range of site-specific biophysical factors; along with the application of a case study to test robustness of the meta-analysis. This study provides an understanding of the effects of management type gradients on coffee yield and whether the low yields in AFSs compared to conventional coffee monocultures, are compensated by the provisioning of ESs.