We report an efficient quantum algorithm for estimating the local density of states (LDOS) on a quantum computer. The LDOS describes the redistribution of energy levels of a quantum system under the influence of a perturbation. Sometimes known as the "strength function" from nuclear spectroscopy experiments, the shape of the LDOS is directly related to the survivial probability of unperturbed eigenstates, and has recently been related to the fidelity decay (or "Loschmidt echo") under imperfect motion-reversal. For quantum systems that can be simulated efficiently on a quantum computer, the LDOS estimation algorithm enables an exponential speed-up over direct classical computation.PACS numbers: 05.45.Mt, 03.67.Lx A major motivation for the physical realization of quantum information processing is the idea, intimated by Feynman, that the dynamics of a wide class of complex quantum systems may be simulated efficiently by these techniques [1]. For a quantum system with Hilbert space size N , an efficient simulation is one that requires only Polylog(N ) gates. This situation should be contrasted with direct simulation on a classical processor, which requires resources growing at least as N 2 . However, complete measurement of the final state on a quantum processor requires O(N 2 ) repetitions of the quantum simulation. Similarly, estimation of the eigenvalue spectrum of a quantum system admitting a Polylog(N ) circuit decomposition requires a phase-estimation circuit that grows as O(N ) [2]. As a result there still remains the important problem of devising methods for the efficient readout of those characteristic properties that are of practical interest in the study of complex quantum systems. In this Letter we introduce an efficient quantum algorithm for estimating, to 1/Polylog(N ) accuracy, the local density of states (LDOS), a quantity of central interest in the description of both many-body and complex fewbody systems. We also determine the class of physical problems for which the LDOS estimation algorithm provides an exponential speed-up over known classical algorithms given this finite accuracy.The LDOS describes the profile of an eigenstate of an unperturbed quantum system over the eigenbasis of perturbed version of the same quantum system. In the context of manybody systems the LDOS was introduced to describe the effect of strong two-particle interactions on the single particle (or single hole) eigenstates [3,4,5,6,7]. More recently, the LDOS has been studied to characterize the effect of imperfections (due to residual interactions between the qubits) in the operation of quantum computers [8,9]. This profile plays a fundamental role also in the analysis of system stability for few-body systems subject to a sudden perturbation [10], such as the onset of an external field, and has been studied extensively in the context of quantum chaos and dynamical localization [11,12]. Quite generally the LDOS is related to the survival probability of the unperturbed eigenstate [4,10,13], and there has been considerable re...