This work traits the continuous-time delays system identification from sampled input-output data focusing on the case when the noise part of the data generating system is an additive white noise. The developed scheme estimates the parameters of the plant and the multiple time-delays separately. More precisely, the parameters are searched applying a Levenberg-Marquardt algorithm while the time delays are estimated by the Gauss-Newton algorithm. Numerical example is presented to show the properties of the proposed scheme.