Linear filter methods have been used in the field of continuous-time identification over a considerable time period. Due to the effectiveness and simplicity of the approach, they have found widespread applications and drawn much interest from the system identification community. However, the estimation of time delay along with continuous-time model parameters has remained an unsolved problem although there are some simple step response based methods. In this paper, a new linear filter method is introduced for simultaneous parameter and delay estimation of continuous-time transfer function models. The proposed method estimates the time delay along with other model parameters in an iterative way through simple linear regression. In addition, the estimated delay is not necessarily an integer multiple of the sampling interval. The applicability of the developed procedure is demonstrated by simulations as well as a laboratory scale experimental example. (C) 2005 Elsevier Ltd. All rights reserved.