Emerging applications of quadrotor vertical take-off and landing (VTOL) unmanned aerial vehicles in various fields have created a need for demanding controllers that are able to counter several challenges, inter alia, nonlinearity, underactuated dynamics, lack of modeling, and uncertainties in the working environment. This study compares and contrasts type-1 and type-2 fuzzy neural networks (T2FNNs) for the trajectory tracking problem of quadrotor VTOL aircraft in terms of their tracking accuracy and control efforts. A realistic trajectory consisting of both straight lines and curvatures for a surveillance operation with minimum snap property, which is feasible regarding input constraints of the quadrotor, is generated to evaluate the proposed controllers. In order to imitate the outdoor noisy and time-varying working conditions, realistic uncertainties, such as wind and gust disturbances, are fed to the real-time experiment in the laboratory environment. Furthermore, a cost function based on the integral of the square of the sliding surface, which gives the optimal parameter update rules, is used to train the consequent part parameters of the T2FNN. Thanks to the learning capability of the proposed controllers, experimental results show the efficiency and efficacy of the learning algorithms that the proposed T2FNN-based controller with the optimal tuning algorithm is 50% superior to a conventional proportional-derivative (PD) controller in terms of control accuracy but requires more control effort. T2FNN structures are also shown to possess better noise reduction property as compared to their type-1 counterparts in the presence of unmodeled noise and disturbances.