To address the problem of performance analysis of sliding-window-accumulation detection (SWA) and order-statistics sliding-window-accumulation detection (OS-SWA) of extended target under space-time correlated K-distributed (C-KD) clutter background, the approximate PDF expressions for the test statistics under null hypothesis were derived on the basis of the theory of generalized K-distribution fitting, moment matching, effective sample size estimation of relevant data and fractional order statistics computing. The relationships between false-alarm probability, CFAR detection threshold and clutter parameters were established. Monte-Carlo test was used to evaluate the precision of the fitted PDF and the false-alarm-rate controlled by the given detection threshold. The effectiveness and reasonableness of the proposed methods has been demonstrated with theoretical analyses and simulation experiments.