This paper describes an analysis procedure for determining the average steady-state mean-square performance of any adaptive multichannel blind deconvolution (MBD) algorithm. The analytical procedure uses the ordinary differential equation corresponding to a linearized version of the algorithm's coefficient updates, from which the mean-square performance can be obtained from the solution of an algebraic Riccati equation. Simulations verify the accuracy of the analysis method as applied to three different multichannel blind deconvolution algorithms.