In 2001, Hsu et al. proposed a non‐repudiable threshold proxy signature with known signers. In their scheme, the proxy group cannot deny having signed the proxy signature if they did. However, Hsu et al.'s scheme is vulnerable to some attacks. A malicious original signer or malicious proxy signer can impersonate some other proxy signers to generate proxy signatures. In this article, we shall present our cryptanalysis of the Hsu et al.'s scheme. After that, we shall propose a new threshold proxy signature that can overcome the weaknesses.