Software reliability stands as a crucial attribute for intricate computing systems, as its absence can lead to a cascade of issues such as increased costs, project delays, and tarnished reputations for software providers. Therefore, ensuring software reliability prior to customer delivery is paramount for any company. Timely error detection, with a reasonable level of accuracy, is crucial for preventing potential consequences. Despite the existence of various software reliability growth models, many of them rely on unrealistic assumptions about development and testing environments, often using black box methodologies. In response to this challenge, a hybrid forecasting model is proposed in this paper. The model combines artificial neural network (ANN) and seasonal auto-regressive integrated moving average (SARIMA) approaches, which are optimised by Jaya optimisation. Improving overall software reliability and software fault predicting are the main objectives. Because it detects possible faults early on, software reliability forecasting is essential to both programme development and maintenance. With Jaya optimisation, the hybrid model combines the complementing advantages of ANN and SARIMA to produce more accurate forecasts and improved flexibility to the intricacies of dynamic software systems. Empirical assessment with real-world software data shows that this hybrid approach outperforms conventional forecasting techniques. The development of more durable and resilient software systems is greatly aided by this research, which is crucial given the quickly changing nature of technology today.