As there is still a great argument on islanding detection methods, this paper presents a new smart passive graphical islanding detection method. The new method depends on measuring the instantaneous variation in phase voltage angle (θ) related to the instantaneous variation in phase voltage magnitude (V) and drawing the relationship between them. The traditional passive methods have performance deficiencies. On the other hand, the active methods are more accurate islanding detection methods, but perturb the electrical system. So, the need for a new detection method which can be accurate, doesn't affect the system behavior, and through which both active and passive methods disadvantages can be avoided is an essential requirement. The continuous updates in Wide Area Monitoring, Protection and Control (WAMPC) technology and Phasor measurement units (PMU) facilitate the idea of measuring and tracing the voltage angle variation. The concept of the new proposed smart passive islanding detection method depends on observing the behavior of the variation of voltage magnitude versus the voltage angle variation (V - θ) of the bus which connects the suggested island with the main power grid. During islanding, sometimes the variation of the voltage magnitude is not significant to be detected, especially in case of low power island. The variation of the voltage angle presents an accurate indication with high sensitivity to the different sorts of system transients. The angle variation behavior differs in load changing, islanding, faulted network, and normal operation conditions. The relationship between (V - θ) is confirmed by the graphical relationship between their derivatives (dV/dt - dθ/dt). V versus θ and dV/dt versus dθ/dt are studied and analyzed in this research to verify the concept of the new method. The new proposed smart method is applied to four different power systems having the same main power grid. The wind electrical power generation unit of the studied power systems has different generation and loading capacities in each of the four studied cases. The studied cases are classified into low, high, and medium (with two different loading status) wind electrical power generation capacity. For each studied case, the system normal operation, load disconnection, islanding, and three phase fault conditions are discussed. In the four studied cases, the results verify the completely different graphical behaviors for each transient condition. So, the concept of the new method is verified.