Centralized cloud storage platforms suffer from inherent limitations such as single points of failure, data privacy concerns, and dependency on trusted service providers. Decen- tralized peer-to-peer (P2P) storage systems aim to overcome these issues by distributing data across multiple independent nodes without centralized control. This paper presents the design and implementation of a decentralized cloud storage system developed in Java using the Kademlia Distributed Hash Table (DHT) for scalable peer discovery and routing. Files are divided into fixed-size chunks, encrypted locally, and distributed across participating peers using content addressing based on SHA- 256 hashes. The system supports decentralized node discovery, encrypted chunk storage, metadata-based file reconstruction, and fault tolerance through replication. Experimental evaluation on a local multi-peer environment demonstrates efficient lookup performance and improved data availability under limited peer churn.