Cloud computing has become a prevalent service for data proprietors to outsource their data to public cloud servers while allowing data consumers to retrieve cloud-stored data. While encrypting cloud data helps individuals ensure the security and privacy of cloud data, all-or-nothing encryption hinders effective access control and data search. To address this issue, this paper proposes fine-grained attribute-based encryption supporting multi-keyword-based data search to circumvent critical issues, including the assumption of online third-party authority, expensive user revocation, and a lack of expressiveness on keyword search problems. The proposed protocol empowers users to authorize cloud servers to perform keyword searches on encrypted data without forfeiting data privacy. Besides, the length of the ciphertext and the user key is short and fixed, having no noteworthy impact on the user growth in the system. The proposed protocol is formally secure against the indistinguishability under chosen-plaintext (IND-CPA) attack under the standard model with the generalized decisional Diffie-Hellman assumption. The comprehensive performance analysis of the proposed scheme demonstrates that it outperforms state-of-the-art solutions. Thus, our system is suitable for real-world applications due to its enhanced security characteristics, adaptability, and efficacy.