Naghibi, A. and Riahi, M.A., 2011. Utilizing modified Gustafson-Kessel algorithm to estimate log data from seismic attributes. Journal of Seismic Exploration, 20: 347-356. One of the most effective methods for analyzing seismic attributes is Fuzzy C-Means (FCM) clustering. By extension of FCM, standard Gustafson-Kessel (G-K) algorithm is derived, which is a powerful tool for clustering analysis. However, G-K algorithm suffers from some shortcomings like singularity of the covariance matrix. By using different techniques for estimating covariance matrix, we can improve the performance of G-K algorithm and lessen the impacts of such pitfalls. Recently, standard G-K algorithm was used for estimation of log data from seismic attributes. In this article we applied the same procedure but instead of standard G-K algorithm we utilized modified G-K algorithm in which we employed two new formulas for further precise estimation of covariance matrix. Because of drawbacks of standard G-K algorithm due to covariance matrix, increasing the number of clusters is not possible in this estimation. Therefore, utilizing recent new techniques have assisted to overcome the drawbacks and improve log data estimation.