Secure multi-party computation (MPC) enables mutually distrusting parties to jointly evaluate a function over their private inputs without revealing those inputs to any participant. Classical MPC constructions based on Boolean or arithmetic circuits incur significant computational overhead when deployed over large participant groups, and their security guarantees rest upon hardness assumptions that are vulnerable to quantum adversaries. This paper introduces an elliptic curve cryptography-based secure MPC framework that exploits the algebraic structure of elliptic curve groups over finite fields to construct efficient oblivious transfer extensions, Beaver triple generation, and verifiable secret sharing protocols. The framework models the computation as a directed acyclic graph over the elliptic curve group law, enabling algebraic optimizations that reduce the number of point multiplications required per gate evaluation. Pedersen commitments over the curve group provide unconditionally hiding and computationally binding commitment guarantees, while zero-knowledge proofs of knowledge of discrete logarithms authenticate each party’s contribution without information leakage. Security is formalized in the universal composability (UC) framework under the decisional Diffie-Hellman (DDH) assumption in the standard model. Evaluation on benchmark circuits including AES key expansion, SHA-3 compression, and private set intersection demonstrates that the proposed EC-MPC protocol achieves 42.3 ms per gate for a four-party evaluation, outperforming BGW by 52.8% and GMW by 42.9% in identical configurations. The framework scales to 32-party computations with sub-quadratic communication overhead, confirming its practicality for cloud-based secure computation, privacy-preserving machine learning inference, and federated healthcare analytics.