The origin question came from I want to move from A to B with a conic curve, and its model was like this:
Say We have a curve transformed from $y=x^2$ (just with rotation and movement)
And is there a transform that could let me calculate the samples of the curve?
in another word I want samples serial to be like this [(5.5,-0.5), ... (1,2)]

Let's start with a parametric form of the curve: $$ \left\{ \begin{array}{ccc} x &=& t \\ y &=& t^2 \end{array} \right. $$ When $t$ gets some range, it plots the usual graph of $y= x^2$. Now we must rotate it about the $z$-axis, with some angle $\theta$: $$ \left[ \begin{array}{c} x_{\text{rotated}} \\ y_{\text{rotated}} \end{array} \right] = \left[ \begin{array}{cc} \cos{\theta} & -\sin{\theta} \\ \sin{\theta} & \cos{\theta} \end{array} \right] \left[ \begin{array}{c} x \\ y \end{array} \right] = \left[ \begin{array}{c} \cos{\theta} t - \sin{\theta}t^2 \\ \sin{\theta}t + \cos{\theta}t^2 \end{array} \right] $$ So if you choose $\theta$, this set of points represents the rotated curve. The only thing left is to displace the curve by a desired amount, let's say $$ \left[ \begin{array}{c} x_{\text{final}} \\ y_{\text{final}} \end{array} \right] = \left[ \begin{array}{c} x_{\text{rotated}} \\ y_{\text{rotated}} \end{array} \right] +\left[ \begin{array}{c} x_{\text{displacement}} \\ y_{\text{displacement}} \end{array} \right] = \left[ \begin{array}{c} \cos{\theta} t - \sin{\theta}t^2 + x_{\text{displacement}} \\ \sin{\theta}t + \cos{\theta}t^2 + y_{\text{displacement}} \end{array} \right] $$ And we're done.