The answer to your question is no.
To clearly see why, let us consider $M=\mathbb{R}^{2}$, $N=\mathbb{R}$, $f\colon (x,y)\;\mapsto\;q=f(x,y)=x$, $X=\frac{\partial}{\partial x} + \frac{\partial}{\partial y}$ and $Y=\frac{\partial}{\partial q}$.
A direct computation shows that $Tf\circ X = Y\circ f$, and it is clear that the flow of $X$ is not constant along the fibre of $f$.
What is important to note, though, is that the flow of $X$ decouples into a part which is purely "horizontal" with respect to $f$ (the flow of $\frac{\partial}{\partial x}$) plus a part which is purely "vertical" with respect to $f$ (the flow of $\frac{\partial}{\partial y}$), and these two flows commute (the decomposition obeys the so-called "principle of composition of independent motions") because $[\frac{\partial }{\partial x},\,\frac{\partial}{\partial y}]=0$.