This work presents method of identification of the functional parameters of a voltage-controlled oscillator (VCO) using artificial neural network (ANN). The VCO under test is excited with a selected stimuli and the ANN is trained by the VCO response in time-domain. Investigated VCO is a mixed-signal circuit, therefore spread of circuit parameters (tolerance) is modelled by Monte-Carlo analysis. Existence of a circuit failures, modelled as spot defects, is also taken into consideration. Additionally, a genetic algorithm (GA) is used to minimize number of circuit response samples. The goal is shortening of data acquisition and processing time during circuit testing, as having significant influence on final manufacturing cost. The proposed method enables shortening test time of the VCO, together with high efficiency of functional parameters identification.