In this paper, a new and powerful metaheuristic optimization technique known as the Coyote Optimization Algorithm (COA) is proposed for the parameter extraction of the PV cell/module. It is utilized to identify the parameters of the single diode and two-diode models. Inspired by the social norms adopted by the coyotes to ensure the survivability of their species, the COA possesses several outstanding merits such as low number of control parameters, ease of implementation and diverse mechanisms for balancing exploration and exploitation. For physically meaningful solutions, a set of parametric constraints is introduced to prevent the coyotes from straying outside of the predefined boundaries of the search space. Extensive tests indicate that the proposed optimizer exhibits superior accuracy compared to other state-of-the-art EA-based parameter extraction methods. It achieved root-mean-square error (RSME) as low as 7.7301E-04 A and 7.3265E-04 A, for the single-diode and two-diode models, respectively. Moreover, the algorithm maintains outstanding performance when tested on an assortment of modules of different technologies (i.e. mono-crystalline, poly-crystalline, and thin film) at varying irradiance and temperature. The standard deviations (STDs) of the fitness values over 35 runs are measured to be less than 1 x 10(-5) for both models. This suggests that the results produced by the algorithm are highly consistent. With these outstanding merits, the COA is envisaged to be a competitive option for the parameter extraction problem of PV cell/module.