The specific recognition of pathogen effectors by intracellular nucleotide-binding domain and leucine-rich repeat receptors (NLRs) is an important component of plant immunity. NLRs have a conserved modular architecture and can be subdivided according to their signaling domain that is mostly a coiled-coil (CC) or a Toll/Interleukin1 receptor (TIR) domain into CNLs and TNLs. Single NLR proteins are often sufficient for both effector recognition and immune activation. However, sometimes, they act in pairs, where two different NLRs are required for disease resistance. Functional studies have revealed that in these cases one NLR of the pair acts as a sensor (sNLR) and one as a helper (hNLR). The genes corresponding to such resistance protein pairs with one-to-one functional co-dependence are clustered, generally with a head-to-head orientation and shared promoter sequences. sNLRs in such functional NLR pairs have additional, non-canonical and highly diverse domains integrated in their conserved modular architecture, which are thought to act as decoys to trap effectors. Recent structure-function studies on the Arabidopsis thaliana TNL pair RRS1/RPS4 and on the rice CNL pairs RGA4/RGA5 and Pik-1/Pik-2 are unraveling how such protein pairs function together. Focusing on these model NLR pairs and other recent examples, this review highlights the distinctive features of NLR pairs and their various fascinating mode of action in pathogen effector perception. We also discuss how these findings on NLR pairs pave the way toward improved plant disease resistance.