Anonymous communication networks are critical infrastructures for protecting user privacy and defending against network surveillance. However, existing anonymous communication networks lack sufficient randomness in path selection, making them susceptible to adversarial prediction and manipulation. Inadequate identity protection further exposes nodes to targeted attacks. Moreover, purely random circuit selection can compromise fairness in long-term node incentives, reducing nodes’ motivation for participation. To address these issues, this paper proposes FRanCS, a blockchain-based Fair and Randomized anonymous circuit selection scheme. Specifically, the scheme employs rolling hash functions and zero-knowledge proofs (ZKPs) to achieve dynamic anonymity in node identity management. At the same time, an exponential key-value random sampling algorithm is designed to simultaneously ensure unpredictability in path selection and fairness in long-term node incentives. Theoretical analysis and experimental results demonstrate that the proposed scheme effectively resists path prediction attacks and node identity exposure risks, achieves fairness in long-term incentives, and exhibits robust security and efficiency in practical deployment scenarios.