The ISI impact factor is widely accepted as a possible measurement of academic journal quality. However, much debate has recently surrounded this use, and several complex alternative journal impact indicators have been reported. To avoid the bias which may be caused by using a single quality indicator, ensemble of multiple indicators is a promising method for producing a more robust quality estimation. In this paper, an approach based on links between journals is proposed for the capturing and fusion of impact indicators. In particular, a number of popular indicators are combined and transformed to fused-links between academic journals, and two distance metrics: Euclidean distance and Manhattan distance are utilised to support the development and analysis of the fused-links. The approach is applied to both supervised and unsupervised learning, in an effort to estimate the impact and therefore the ranking of journals. Results of systematic experimental evaluation demonstrate that by exploiting the fused-links, simple algorithms such as K-Nearest Neighbours and K-means can perform as well as advanced techniques like support vector machines, in terms of accuracy and within-1 accuracy, while exhibiting the advantage of being more intuitive and interpretable.