Increasingly often, we need to learn predictive models from big or complex data, which may comprise many examples and many input/output dimensions. When more than one target variable has to be predicted, we talk about multi-target prediction. Predictive modeling problems may also be complex in other ways, e.g., they may involve incompletely/partially labelled data, as in semi-supervised learning. The talk will first give an introduction to the different tasks of multi-target prediction, such as multi-target classification and regression, hierarchical versions thereof, and versions of the tasks that involve additional complexity (such as semi-supervised multi-target regression). It will continue to present some methods for solving such tasks. Finally, it will review different applications of multi-target prediction in environmental and life sciences, ranging from relating environmental conditions and the composition of biota, through gene function prediction, to predictive modeling in virtual compound screening for drug repurposing