Long Term Evolution (LTE) is a promising positioning system for its appropriate positioning characteristics and excellent coverage, especially in the harsh environments such as urban canyons and indoor scenarios. However, most of the researches so far on LTE positioning algorithms are based on time parameters. Thus in this paper, an algorithm based on signal parameters estimation via rotational invariance techniques (ESPRIT) to jointly estimate the time-of- arrival (TOA) and the direction-of-arrival (DOA) from received LTE signals in multipath environment is presented. This method estimates the TOA parameters first after modeling the received LTE signals in the frequency domain. Then the DOA parameters via the difference of the TOA are obtained, and thereby the joint TOA and DOA estimation is gained. Simulation results show that this method performs well on different signal-to-noise-ratios, numbers of multipath and transmission bandwidths. Moreover, the simulation results show that the proposed algorithm has much better performance than multiple signal classification (MUSIC) algorithm and propagator method (PM) and its positioning accuracy can achieve centimeter level.