<p>Boost converters are used in many applications and most common in the photovoltaic (PV) systems within DC microgrids. Nevertheless, the open-loop control-to-output voltage transfer function (CVTF) of this converter may show non-minimum phase behavior which may impose adverse effects on the control and stability of the PV system and thus, the DC microgrid. Accordingly, identification of the boost converter of the PV systems considering the non-minimum phase effects can help recognizing the control and stability problems. Therefore, this paper first derives the CVTF of the boost converter analytically through small- and large-signal disturbance analysis. Then, characteristics of the CVTF are investigated through time domain solution to the differential equations, frequency response analysis, and simulation-base sensitivity analysis to the duty cycle noise. After that, two techniques are used to identify the CVTF model: The first technique uses time domain data and considers the boost converter model as a stochastic process model. This technique uses Auto-Regressive with exogenous Input (ARXI) method to decompose the linear and nonlinear parts of the boost converter model and minimize the estimation errors using the QR decomposition without regularization. The second technique uses the discrete frequency domain data and adopts a new proposed Instrumental Variable-Approximate Maximum Likelihood Method (IVAMLM) to identify the converter model. The frequency domain data are generated through successive simulations of a PV system by exciting the model through a variable frequency sine-stream wave. The comparison results of the new IVAMLM with those founded with the ARXI show the feasibility of the estimated models by investigating various estimation criteria and under-fitting and over-fitting issues.</p>