Decide what you want to portray in the music video. If you watch various music videos you will see that some (most) have a mixture of 'band singing/playing' footage and a separate story to help convey the message or mood of that song.
Your first step is to come up with the message that you want your video to express. What does your song convey? Once you have that determined then you can start planning out how you will get that across visually.
For example, say you have selected a country song. The song is about young love and infatuation. What do you see in your mind when you listen to this song? It may be of two young people madly in love who are enjoying being youthful while they sit on a truck-bed gazing at the stars, holding each other in their arms, etc. It may be a montage of them just being together running through the praire fields, making out in a pile of hay, etc. I dunno, this is just an example. haha
Once you have that, you just cast and go out and film it!
If you want a mixture of band members singing the song then you just film them and cut it into the music video.
Sorry if this is confusing, I wrote it in a hurry.
Hope this helps!