Subscriber churn prediction is a critical challenge for businesses operating across multiple domains such as telecommunications, banking, and subscription-based services. Accurately identifying customers who are likely to discontinue services enables organizations to implement proactive retention strategies. This study proposes a multi-domain churn prediction framework using Artificial Neural Networks (ANN) to improve predictive performance across diverse datasets The model integrates heterogeneous data sources, including customer demographics, usage patterns, transaction history, and service interactions, to capture domain-specific and cross-domain behavioural patterns. A unified ANN architecture is designed and trained on combined datasets, allowing the model to generalize across domains while preserving domain-level distinctions. Feature engineering and normalization techniques are applied to ensure consistency and improve learning efficiency.