Once a central topic of philosophy of science, scientific explanation attracted less attention in the last two decades. My aim in this paper is to argue for a new sort of approach towards scientific explanation. In a first step I propose a classification of different approaches through a set of dichotomic characteristics. Taken into account the tendencies in actual philosophy of science I see a local, dynamic and non-theory driven approach as a plausible one. Considering models as bearers of explanations will provide a proper frame for such an approach. In the second part I make some suggestions for a working agenda that will further articulate a sketchy account of explanation through models proposed by Hartmann and Frigg.