Abstract: Many application fields require detecting real-time composite events and reacting to them immediately considering their sensitive condition. This raises the need for efficient, situational aware Complex Event Processing (CEP) engines to handle massive amount of streamed data. To satisfy this need an optimization method to enhance the engine's efficiency and reliability is proposed. It adopts a combination of enhanced buffer allocation method with pattern matching process under the supervision of domain expert. An error detection and recovery to handle any failure case is proposed with the system to assure high reliability and accuracy in delivering and detecting the events. The proposed system is evaluated in terms of performance over time, throughput, error rate and pattern detection accuracy. A comparative evaluation with a state-of-art engine is done and the proposed system outperformed it under challenging workloads.