Despite the reduction of solar installation cost by a ratio of 10 during the last 20 years, still conversion efficiency of PV panels remains an important parameter to take into account when designing solar system for residential applications [1]. Therefore, an efficient MPPT algorithm plays a key role to harvest optimum available power especially in large solar installations. Several MPPT algorithms have been elaborated in the literature. Among them, perturb and observe (P&O) and incremental conductance (INC) are practically favorable as the awareness of PV panel characteristics is not required. Moreover, their procedures to find MPP are independent of temperature and irradiation values that their measurements need expensive sensors [2,3]. Principally, P&O and INC are inherently perturbative methods that produce reference voltage or current to be tracked by a subsequent controller. The PI controller has been widely used in MPPT operations [4,5]; however, continuous evolution in microprocessor technology facilitates implementing advanced controllers for efficiency enhancement of MPPT algorithm. The ability of Fuzzy logic [6,7], neural network [8] and genetic algorithm [9] has been investigated in MPPT modules. This paper deals with predictive technique that is lately well adopted for various applications. Indeed, model predictive controller (MPC) is a competitive alternative to address the growing industrial concerns regarding to performance and efficiency issues. It can also formulate inherent nonlinearity in power electronic systems with operational constraints. Moreover, its realization in state matrix can be easily extended to multivariable systems [10]. Basically, MPC solves an optimization problem within a moving time horizon in order to generate future actions for optimal operation of a plant. In fact, at each sampling time, MPC reconstructs instant operating model of the plant, predicts future states and optimizes current dynamic while taking into account the future states. Real time modification is a desirable property in practice that can compensate inevitable modeling errors [11]. In power electronics, MPC were emerged with finite control set (FCS) appearance. Actually the pulse activation nature of switching converters allowed defining FCS-MPC that evaluates cost function in only possible switching states. A low complexity optimization algorithm is solved and simply the minimum cost is selected among predetermined states at each sampling time. FCS-MPC has been investigated in most applications of power converters [12–۱۴], and especially in PV systems [15–۱۸]. Despite numerous reports in this area, the applicability of FCS-MPC is yet a challenge. It works with variable switching frequency which leads to a widespread harmonics spectrum for voltage/current waveforms. This is a fundamental limitation that hinders filter design [19,20], increases switching losses, makes unwanted resonances, and consequently reduces system performance in terms of power quality [21]. However, the switching frequency could be regulated by including some more terms in the cost function, it would add the complexity and consequently the computation burden to the FCS-MPCs while distracting control effort from main target of the reference tracking [22,23]. Moreover, none-zero steady state error is reported in [24,25]. Furthermore, the computational burden increases exponentially in multilevel converters with high number of switching states [26].