This paper presents a new strategy for estimation of state variables. The strategy may be applied to linear systems and is referred to as the Variable Structure Filter. The filter is considered for discrete-time systems subject to random disturbances and measurement noise. A proof of stability for the filter is provided. For stability, this filtering concept requires a specification of an upper bound for the random disturbances and measurement noise. The filter requires a parametric model and can be formulated to accommodate model uncertainties. The application of this filter to a third order linear system is demonstrated.