A number of aldose reductase inhibitors (ARIs) have been developed over the past few decades with the expectation of therapeutic effects for diabetic complications. Neuropathy is the complication that has been most intensively studied as a potential target for ARIs. Most ARIs have shown satisfactory effects in animal models. However, the clinical potential of ARIs in diabetic patients has been controversial due to the lack of conclusive evidence. The safety of this category of drugs is also uncertain. This article summarizes the results of clinical trials of ARIs for patients with diabetic neuropathy that have been performed to date. The efficacy and toxicity of each ARI will be briefly assessed by the clinical data. The theoretical background along with major issues in the evaluation of drug efficacy will also be discussed. Overall the observed efficacy varied among the compounds. A few ARIs showed favorable effects in multiple endpoints in the majority of trials, while the results from many ARIs seemed ambivalent. One drug barely exhibited positive effects on any endpoint. This discrepancy may be attributable at least in part to the different degree of inhibition of the polyol pathway in nerve tissues, which is determined not only by the pharmacokinetic properties of the drug but also by its penetration into nerve tissues. In addition to the uncertain potential of each ARI, the issues of design and analytical methods used for clinical trials may underlie the ambivalent outcomes. The power of analysis and the duration of trials were apparently inadequate in a large number of the studies. Various indices selected as endpoints are not necessarily sensitive or reproducible. Studies of longer duration, large-scale trials, better methods to assess neuropathy, and the selection of patients with a homogenous background would provide more conclusive evidence. The risk of serious adverse reactions, for example, hypersensitivity reactions and hepatic damage, has led to some ARIs being withdrawn from the market or from further development. These adverse effects, however, do not appear to result from the inhibition of aldose reductase activity per se but from specific reactions to each compound. In conclusion, sufficient inhibition of the nerve aldose reductase activity seems likely to prevent or ameliorate diabetic neuropathy, and further development of more potent and safe ARIs is necessary before extensive clinical application.