In this paper, the problem of l2ā€“lāˆž mode-independent filter design has been investigated for discrete-time singular Markovian jump systems with state delays. A sufficient condition has been derived for the filtering error system to be regular, causal, and delay-dependent stochastically stable. Based on the obtained analysis result, the full-order and reduced-order filters have been designed in a unified framework. The addressed filter design problem has been converted into the feasibility problem of a set of LMIs, which can be checked efficiently via the Matlab LMI Toolbox. The obtained delay-dependent results rely upon the partitioning size. The numerical examples have been given to show the effectiveness and the potential of the proposed techniques.