This paper presents a recursive algorithm using orthogonal block-pulse functions to identify the parameters of a system described by a linear differential equation with coefficients of finite-order polynomials in time t. The key point is the derivation of a set of operational matrices Pij(i = 0, 1, 2, … ; j=0, 1, 2, …) that relate the (th integral of the time-multiplicated block-pulse vector tiφ(t) to the block-pulse vector φ(t) itself. By using this elegant set of operational matrices the original time-varying differential input-output model is converted to a set of linear algebraic equations which can be solved recursively by a least-squares method to give unknown parameters and initial conditions. Two examples are included, one to illustrate the superiority of Pijover Fj(Wang 1982) in the solution of a time-varying differential equation and the other to show the effectiveness of the identification method.