Three JNK isoforms, JNK1, JNK2, and JNK3 have been reported and unique biological function has been ascribed to each. It is unknown if selective inhibition of these isoforms would confer therapeutic or safety benefit. To probe JNK isoform function we designed JNK2/3 inhibitors that have >30-fold selectivity over JNK1. Utilizing site-directed mutagenesis and x-ray crystallography we identified L144 in JNK3 as a key residue for selectivity. To test whether JNK2/3 selective inhibitors protect human dopaminergic neurons against neurotoxin-induced mitochondrial dysfunction, we monitored reactive oxygen species (ROS) generation and mitochondrial membrane potential (MMP). The results showed that JNK2/3 selective inhibitors protected against 6-hydroxydopamine-induced ROS generation and MMP depolarization. These results suggest that it was possible to develop JNK2/3 selective inhibitors and that residues in hydrophobic pocket I were responsible for selectivity. Moreover, the findings also suggest that inhibition of JNK2/3 likely contributed to protecting mitochondrial function and prevented ultimate cell death.