Yes it is.
Just one quick example: often physicists have to solve differential equations, like [tex]\mathcal{L}\phi = f[/tex] for given [tex]f[/tex] and differential operator [tex]\mathcal{L}[/tex]. One way do do this is by constructing a Greens function which satisfies [tex]\mathcal{L}G(x) = \delta(x)[/tex] (that's a Dirac delta) and then the equation can be solved for any [tex]f[/tex] by convolution; [tex]\phi = G * f = \int G(x - x') f(x') dx'[/tex]. Once you do this for different differential operators, you'll notice that it's often much handier to solve the Fourier components of [tex]G[/tex] separately (especially since the delta function has such an easy Fourier transform) and then back-transform them to get G.
In fact, I have heard that in field theories (QFT, for example) people love working in Fourier space, as problems are often relatively simple there and a real pain in the neck to do in real space.
So my advise would be: if you can study Fourier analysis, do it.