This Sensor faults have been a major area of research for fault tolerant system. Fault accommodation system mainly deals with two specific challenges. First, being an efficient detection of fault and second, a robust model whose estimated sensor signal may be used to reconfigure the system in case of fault. Research shows advancement in model based techniques for fault accommodation. An expert system is presented in this paper with integration oftwo neural network based approaches to provide a robust fault accommodation system. A dedicated knowledge based neural network is used for fault detection only. This detection system is integrated with another model based neural network which provides sensor estimate. Hard sensor stuck fault and intermittent fault for F-8C aircraft and jet transport aircraft is simulated to show the proposed fault accommodation system. The proposed method is tested on closed loop model of the aircraft and validated with ARM target, i.e. SAM3X8E. The ADC and DAC ports of the board were used for interfacing purpose. The input command of pilot and the cockpit display was interfaced with ARM Cortex-M3 CPU.