You can use GDI+ for doing this. There is also
this[
^] nice article on CP which might help you.
Update: Looks like we all missed the point. Assuming you need to draw a cube or a cuboid, length, breadth and height are not enough. This will never tell you the intended orientation of the object in the space. If that is all the information you have, you will end up drawing your perception of object in space. I hope I am clear. To get the real position in the space of the object, you will need at least some co-ordinates if not all. To me it is not enough input to draw the object. You can very well draw the isometric view and then let the user rotate the object. This, surely, will need you to handle the axis about which the rotation is done and also provide a smooth flicker-free movement. I am sure there are libraries available that ca help you do this. So you might be better off using a readily available component rather tht doing yourself. Although it's your call.
Another thing that I noticed is use of "vector". Are you referring that those three input numbers will be the vector values in i, j and k axis and then you need to draw a projection of that? This is totally different from drawing a rectangle. This will involve you drawing upto 4 triangles who share sides with each other. This again will be easier to do using the already available libraries than doing yourself. Your call again.