The cloud provides convenient storage services for the vast amount of data generated by the Internet of Things (IoT), yet it also introduces security challenges such as data tampering and privacy breaches to IoT. While IoT encrypts sensitive data before storing it on cloud servers, it is unable to perform searches on encrypted data through these cloud servers. The advent of searchable encryption technology has successfully solved the problem of searching encrypted data, thereby protecting user data privacy to a certain extent. Most existing searchable encryption schemes support precise keyword searches, with the cloud servers providing encrypted search services. However, cloud servers may perform partial searches or fabricate some results due to various incentives, such as saving computational or storage resources. To address the aforementioned issues, we proposed BC-RFMS, a blockchain-based rankable fuzzy multi-keyword search scheme. Leveraging the immutable properties of blockchain, BC-RFMS ensured the accuracy of search results through smart contracts for fuzzy searching. Furthermore, the scheme utilized locality-sensitive hashing and Bloom filters to construct fuzzy keyword search indices, while employing term frequency-inverse document frequency (TF-IDF) for relevance score computation and ranking. We also proposed a block-by-block transfer algorithm to prevent surpassing the GasLimit when uploading data to the blockchain. To enhance storage security and reduce Gas consumption, encrypted data was stored on the InterPlanetary File System (IPFS) for distributed storage. Experimental analyses conducted on a blockchain test network validated the feasibility of the BC-RFMS scheme.