In this paper, a Harmony Search-based Hyper-heuristic (HSHH) approach is proposed for tackling examination timetabling problems. In this approach, the harmony search algorithm will operate as a high level of abstraction which intelligently evolves a sequence of low level heuristics. This sequence is a combination of improvement heuristics which consist of neighborhood structure strategies. The proposed approach is tested using the examination timetabling tracks in Second International Timetabling Competition (ITC-2007) benchmarks. Experimentally, the HSHH approach can achieve comparable results with the comparative methods in the literature. © Springer International Publishing Switzerland 2014.