The biological processes of waste-water treatment have often been regarded in practice as largely self-controlled and somewhat inflexible in their operation. The paper reviews some of the reasons why it has been difficult to apply the methods of control theory to these processes, and what the future role of control in waste-water treatment might be. Attention is restricted to three processes: the activated sludge process, the oxidation ditch and anaerobic digestion. In contrast to many biotechnical processes, these processes are characterised by a heterogeneous culture of micro-organisms and by multiple substrates degraded along several reaction pathways. Their dynamics are extremely difficult to identify, their observed input/output relationships frequently exhibiting substantial stability punctuated by abrupt instabilities. Other aspects of their behaviour are relatively well defined, in particular the dynamics of dissolved oxygen in the activated sludge process, and have been equally well studied from the point of view of online state-parameter estimation and adaptive control. There is still, however, a large gap between practice and theory, and, in looking briefly at an agenda of problems for the future, the paper places special emphasis on,inter alia, issues of fault detection and diagnosis and the use of plant operator experience in generating novel approaches to control.