Secure multi-party computation is covered in full under multi-party computation, which explains the protocols, history and applications. This page explains the longer name and its abbreviations, sMPC or SMPC, which are common in academic papers and vendor material.
Secure multi-party computation is a branch of cryptography that lets a group of parties calculate a function of their combined data while each keeps its own input secret. A classic illustration is several employees working out their average salary without anyone learning another's pay. The word "secure" stresses the guarantee that nothing beyond the final result is revealed, even if some participants try to cheat, up to a limit set by the protocol.
The field began with Andrew Yao's work in the early 1980s, including his "millionaires' problem" of two people learning who is richer without disclosing their wealth. Later protocols use techniques such as secret sharing, a method related to Shamir's secret sharing, and garbled circuits.
In finance and crypto, sMPC is best known for key management. An MPC wallet splits signing power across several devices or parties, so that a private key never exists in one place and a transaction is signed only when enough parties cooperate. Banks and researchers also use it for privacy-preserving fraud analysis and data sharing. It differs from fully homomorphic encryption, which lets a single party compute on encrypted data.