Artificial Intelligence algorithms predict the future based on the trained models and datasets. However, a reliable prediction requires a tamper-resistant model with immutable data. Blockchain technology provides trusted output with consensus-based transactions and an immutable distributed ledger. Therefore, blockchain can help AI to produce immutable models for trustworthy prediction. But most smart contracts that define the language of blockchain applications do not support floating-point data type, limiting computations for classification, which affects the prediction accuracy. In this work, we propose a novel method to produce floating-point equivalent probability estimation to classify labels on-chain with a Naive Bayes algorithm. We derive a mathematical model with Taylor series expansion to compute the ratio of the posterior probability of classes to classify labels using integers. Our derived method is platform-agnostic to support various blockchain networks. Furthermore, our solution is reproducible for deep-learning algorithms. In the future, we plan to expand our work to support more AI algorithms. We will scale our solution for real-time and inexpensive object classification. Additionally, we plan to develop privacy-preserving AI models using blockchain smart contracts.