The paper deals with dynamic compensation of delayed Self Powered Flux Detectors (SPFDs) using discrete time ${{\bf H}_\infty} $ filtering method for improving the response of SPFDs with significant delayed components such as Platinum and Vanadium SPFD. We also present a comparative study between the Linear Matrix Inequality (LMI) based ${{\bf H}_\infty} $ filtering and Algebraic Riccati Equation (ARE) based Kalman filtering methods with respect to their delay compensation capabilities. Finally an improved recursive ${{\bf H}_\infty} $ filter based on the adaptive fading memory technique is proposed which provides an improved performance over existing methods. The existing delay compensation algorithms do not account for the rate of change in the signal for determining the filter gain and therefore add significant noise during the delay compensation process. The proposed adaptive fading memory ${{\bf H}_\infty} $ filter minimizes the overall noise very effectively at the same time keeps the response time at minimum values. The recursive algorithm is easy to implement in real time as compared to the LMI (or ARE) based solutions.