In this paper, a new symbol timing recovery is proposed that is suitable for burst transmission scheme. It is assumed that the channel is an Additive White Gaussian Noise (AWGN) one and the transmitted data are unknown at the receiver. The contribution of the paper is exploiting Gardner Timing Error Detection (TED) algorithm for None-Data-Aided (NDA) timing delay estimation in an iterative manner. It is shown that the estimated delay also maximizes the log-likelihood function. The algorithm is suitable for different types of linear modulations such as PSK, QAM. The proposed timing recovery algorithm is bandwidth efficient which does not use any pilot symbols and proceeds without using preambles. The algorithm is capable of recovering all symbols of a short burst transmission and is practically easy to be implemented. Simulation results confirm the algorithm convergence to the maximum of the likelihood function and a good performance in terms of Mean Square Error (MSE).