Yes, this is basically the correct approach. Strictly speaking, some of the energy you put in will go into heating the water, i.e., the water will get hotter as you agitate it. The energy balance would be something like:
$$ \Delta E = W + C\Delta T $$
where $C$ is the specific heat and $\Delta T$ the temperature rise. However, under most conditions, the temperature rise will be small enough that it can be ignored.
The work done to keep the helicopter hovering is that generated by the motion of the rotor blades as they move through the air. The two key notions that go into the calculation of this work are lift and drag. Lift is the upwards-pointing force that balances gravity and keeps the helicopter hovering. Drag is the force pushing back horizontally against the rotors in the direction opposite to their motion. Lift is the "useful" force that helicopters and other aircaft need in order to fly, and drag is the unwanted force that is the price we have to pay to get lift.
In the hovering scenario, all the work goes into resisting the drag. If the drag force is equal to $D$, the work will be obtained by multiplying $D$ by the distance the rotor blades moved. We can encode this in the equation
$$
P = \left(2\pi \frac{A}{2}\right) \times f \times D,
$$
where
\begin{align*}
P &= \textrm{power (work per second) to keep the helicopter hovering}
\\
A &= \textrm{length (radius) of the rotors}
\\
f &= \textrm{frequency of rotation of the helicopter's rotors (in revolutions/second)},
\\
D &= \textrm{drag}
\end{align*}
(for obvious reasons, I find it easier to talk about power output in Watts rather than total energy in Joules). The meaning of the $2\pi A/2$ factor is that this is the circumference of a circle whose radius lies along the midpoint of the rotor blades - I'm assuming (which seems reasonable as an approximation) that drag is distributed uniformly along the length of the blades, so that would be the distance we would want to multiply by to get the work for a single revolution of the rotors.
Now, $A$ and $f$ are parameters whose values are easy to figure out (see the numerical example below), but how do we compute $D$? Well, recall that the helicopter is hovering, which means that the lift force $L$ is exactly equal to the weight of the helicopter:
$$
L = \textrm{helicopter weight} = \textrm{(gravitational constant)} \times \textrm{(helicopter mass)}.
$$
The last thing we need is the helicopter's lift to drag ratio, which will tell us how much drag will develop for a given amount of lift. This is a number that I think cannot be calculated from first principles, but must be measured (or approximated using numerical simulations). If the lift-to-drag ratio is $\gamma$, that simply means that $L$ and $D$ satisfy the relation
$$
L = \gamma D.
$$
This enables us to write down our final formula for the power output needed to keep the craft hovering:
$$
P = \pi A f L/\gamma
$$
A numerical example:
According to this Reddit thread, one of the most common helicopters used in civil aviation is the Robinson R22. From the specs on the Wikipedia page for this helicopter, its rotor length is
$$A = 3.83 \ \textrm{m}
$$
(they give the rotor diameter which is twice this value). The lift-to-drag ratio would vary depending on various complicated factors (see here), but according to the Wikipedia article on lift-to-drag ratios, a typical value for a helicopter is
$$
\gamma = 4,
$$
so I'll work with that.
Again referring to the helicopter specs on Wikipedia, the weight of the helicopter would be between 417 and 622 kilograms. Let's assume a value of 500 kilograms. The lift force is therefore equal to
$$
L = 9.8 \times 500 = 4900 \textrm{ N}.
$$
Since $\gamma=4$, the drag is therefore
$$
D = 4900/4 = 1225 \textrm{ N}.
$$
The final parameter is $f$, the number of revolutions per second. According to this source, a typical rate of rotor rotation for a helicopter is between 250 and 600 revolutions per minute. Let's assume 400 rpm, which translates to
$$
f = 400/60 = 6.66 \textrm{ revolutions per second}.
$$
We can now plug all these numerical parameters into the formula above to get the power output. The result is
$$
P = 3.14159 \times 6.66 \times 3.83 \times 1225 = 98165 \textrm{ Watt}
= 133.46 \textrm{ metric horsepower},
$$
i.e., the helicopter needs to do 98 kJ of work per second to maintain hovering.
Are these numbers reasonable? Well, according to the Wikipedia page for the R22, this helicopter uses a powerplant with a power output of 124 horsepower, which is less than my number (even with the unrealistic assumption of 100% engine efficiency). It looks like my numbers are a bit off, but it's not bad for an order of magnitude calculation.
Best Answer
Both seem to be correct, as another answer also points out. Meaning, both formulas can be used to calculate power.
But the fundamental definition of power is the latter, namely energy per time.
In this general definition, power can be calculated for mechanical systems, thermodynamic systems, electric system etc. The former version includes force and might thus be harder to use in, say, thermodynamic problems.