The authors derive a measure suitable for identifying sensor placements that yield high target classification rates in a distributed radar environment. A distance measure is derived as an approximation lower bound of the Kullback–Leibler (KL) divergence between two probable target classes. A relationship between the KL divergence approximation and probability of correct classification (PCC) is demonstrated and used to identify sensor placements likely to yield higher PCC values. Two algorithms are proposed, optimal and approximate, for identifying good placements of a single transmitter and two receivers. A physics-based scattering model is employed to generate reflectivity data suitable for multi-sensor, multi-static classification analysis. KL and PCC data sets are analysed to demonstrate the effectiveness of the proposed distance measure and algorithms.