The development of effective acute stroke therapies is a high priority because of the substantial death and long term disability that result from acute stroke. Reducing the initial extent of brain damage is key if outcomes from acute stroke are to be improved. In 85-90% of cases stroke is due to infarction usually secondary to atherothrombotic or cardioembolic disease causing vessel occlusion in the anterior circulation. The two main drug targets in ischaemic stroke are lysis of the acute thrombus in the target vessel and protection of the ischaemic penumbra. Drug development of thrombolytic therapy has successfully occurred although dose selection to optimise the risk/benefit ratio (haemorrhage/ recanalisation) has proved a difficult challenge. The efficacy of thrombolysis has been clearly demonstrated within 3 hours and translation of animal studies was achieved in proof of concept studies in man examining recanalisation in acute stroke, then efficacy in phase III studies. Phase IV studies have an important role in defining safety in routine clinical practice and in patient subgroups such as the very elderly, poorly represented in phase III clinical trials. Translating neuroprotective drug effects seen in animal models into positive phase II/III studies has been unsuccessful despite considerable industry efforts. Early studies had many problems in study design with failure to achieve plasma concentrations consistent with animal studies, patient selection with lack of initial imaging, and long time windows. The 1999 and 2001 STAIR consensus criteria listed preclinical and phase II/III design requirements intended to improve chances of demonstrating efficacy. Despite improvements in clinical study design the lack of a ‘proof of concept’ model to demonstrate salvage of the ischaemic penumbra in phase II studies remains a major barrier. The recent failure to demonstrate efficacy of the free radical scavenger NXY-059 following an initial positive phase III trial and achievement of most of the STAIR criteria place a question mark over the future of neuroprotective drug development. However demonstration of the benefits of hypothermia following cardiac arrest indicate neuroprotection is achievable in man. From a clinical pharmacological perspective a number of potentially correctable problems in the development of neuroprotection remain including: increasing quality, reporting and analysis of pre-clinical studies; efficacy in animal models more reflective of the older stroke patient with physiological derangement; efficacy in in-vitro human studies (cell culture, tissue slices); demonstration of drug access through the blood brain barrier to the putative site of action (PET, intra-operative studies); selection of patients with salvageable tissue (MR, CT perfusion selection); very early treatment (pre-hospital paramedic based trials); adaptive study design to identify most promising doses for phase III studies; standardisation and refinement of clinical measures of neurological impairment and disability; and physiological optimisation in proof of concept human studies. There remain concerns that rodent acute stroke studies are a poor model of the human ischaemic penumbra. Further refinement and validation of rodent models, and consideration of the role of primate models is required. Despite many difficulties and disappointments neuroprotection in stroke remains a tantalising, achievable goal and one of the great challenges of translation for the next decade.