Digital marketing has revolutionized e-coupon systems, creating an urgent need for secure, transparent, and efficient e-coupon solutions. Current coupon platforms are plagued by issues like forgery, personalization challenges, privacy, and scalability bottlenecks. This paper proposes a novel blockchain-based e-coupon framework leveraging hybrid time series forecasting and privacy-preserving analytics. It incorporates a hybrid forecasting model (ARIMA, VAR, TCN, and LSTM) to dynamically optimize coupon parameters like discount rates, validity period, and target user segments based on transaction data recorded on the blockchain. It also uses zero-knowledge proofs (zk-SNARKs) to guarantee secure and anonymous coupon redemption. In addition, a lightweight consensus mechanism, such as PoS, is utilized to improve the scalability and reduce environmental footprints over the traditional proof-of-work. The proposed system also includes a cross-platform interoperability layer to facilitate the cross-platform redeeming of coupons. The experimental results show improved coupon redemption rates, system throughput, and privacy protection. The proposed solution provides a novel, intelligent, and privacy-centric method for digital coupon management and enables more secure and personalized marketing in the decentralized e-commerce.