The nucleotide-binding domain (NBD) and leucine-rich repeat (LRR)-containing receptors (NLRs) mediate innate immunity by forming inflammasomes. Activation of the NLR protein NLRP1 requires auto-cleavage within its FIIND domain1-7. In resting cells, the dipeptidyl peptidase DPP9 interacts with NLRP1-FIIND and together with a related enzyme DPP8, suppresses spontaneous NLRP1 activation8,9. The mechanisms of DPP8/9-mediated NLRP1 inhibition, however, remain elusive. Here we provide structural and biochemical evidence demonstrating that rat NLRP1 (rNLRP1) interacts with rDPP9 in a stepwise manner to form a 2:1 complex. An auto-inhibited rNLRP1 molecule first interacts with rDPP9 via its ZU5 domain. This 1:1 rNLRP1-rDPP9 complex then captures the UPA domain of a second rNLRP1 molecule via a UPA-interacting site on DPP9 and dimeric UPA-UPA interactions with the first rNLRP1. The 2:1 rNLRP1-rDPP9 complex prevents NLRP1 UPA-mediated higher order oligomerization and maintains NLRP1 in the auto-inhibited state. Structure-guided biochemical and functional assays show that both NLRP1-binding and its enzymatic activity are required for DPP9 to suppress NLRP1, supporting guard-type activation of the NLR. Together, our data reveal the mechanism of DPP9-mediated inhibition of NLRP1 and shed light on activation of the NLRP1 inflammasome.