I think your confusion is due to the fact that there are two types of friction.
- One is always zero. That is kinetic friction (it only appears when something slides, which isn't the case for rolling).
- The other is not always zero. That is static friction (it appears to prevent sliding, so it holds on to the contact point whenever there is a tendency to slide).
When you accelerate or brake with your car, the wheels rolling over the surface feel a static friction force. If not, then the wheels would slide over the surface, because they are being slowed down by the engine (via the axle) but nothing is slowing down the car as a whole and so, the wheels will slide over the asphalt.
Also, as in your example, if you roll anything uphill or downhill, gravity pulls downwards and tries to make the contact point slide. Static friction then appears to prevent that sliding. It must thus pull up along the incline (regardless of rolling direction).
I have one more doubt that what if any external force is present in case of rolling down for rough and smooth inclined plane? I believe that in smooth incline plane the force will provide the torque and the object will roll down but what in case of rough inclined plane.
An external force could for instance be gravity, I guess, unless you are thinking of something specific. But gravity sums up to be pulling in the centre of the wheel. It causes no rotational torque about the centre. So, gravity can't be the thing that causes the rotation.
Rather, static friction at the contact point makes the wheel rotate.
- Is the incline completely smooth, then no static friction is possible. Then no rotation happens and the wheel will simply slide down the incline without ever beginning to spin around.
- Is the incline rough, then a static friction force is possible, and now this static friction causes a torque that makes the wheel rotate.
In other words, a rough surface is a requirement for pure rolling (or rolling without slipping/sliding) to be possible in such scenario.