The sparseness-controlled improved proportionate normalized least mean square (SC-IPNLMS) algorithm can improve the convergence speed for identification of sparse systems as compared to their conventional couterparts. nevertheless, the requirements of past convergence and low seady state misalignment are conflict for constant step-size adaptive algorithms, whose step size parameter has to be selected by compromising bthese two conflict requirements. in this paper, a novel algorithm exploiting a variable step size, named VSS-SC-IPNLMS algorithm is proposed. the proposed algorithm obtained new optimization criterion by forcing the posterior error to cancel negative effect to disturbance signal. then using the relation between posteriori estimation error and priori estimation error, a step size control approach for proportionate algorithm is provided. echo cancellation simulation result confirm that the proposed algorithm can constitute a significant improvement in the convergence speed with verry small miss-adjusment when compared with the constant step-size SC-IPNLMS algorithm and other variable step-size algorithms.