I'm not sure I follow what you mean when you say you want to use your multispectral image as panchromatic. What exactly do you plan to do with it, that requires it to be panchromatic? The average of the three bands would not be all that similar to a true panchromatic image, which would have a signal integrated over the full wavelength range of the panchromatic detector. But maybe that doesn't matter, depending on what you're planning to do with it? Any result from Band Math will not have wavelength info associated with it. So, ENVI won't be considering the wavelengths in anything that it does. So, not panchromatic, and definitely not multispectral.

Thank you Whammer for your respond. I was asked to do a pan sharpening with QB which is multispectral and Landsat (multispectral) also ..... I was told to take one band or a band math(like the average of the RGB bands) from the QB multi and use it as panchromatic.(spatial analysis 0.6x0.6) The Landsat is (28.5x28.5).I am wondering if it is possible..

Hmm. Well that would probably not be the best way to proceed. The pan sharpening algorithms assume that you have a real panchromatic image, which really represent the signal across the wavelength range of the multispectral image you're trying to sharpen. If you do what you are suggesting, that assumption will be violated. But technically it is possible to do what you describe. Go ahead and make the average image using Band Math. Then when you're doing the pan sharpening, just choose that image as your high resolution panchromatic image. I wonder why you don't use the panchromatic QB image that corresponds to your multispectral QB scene? Or better yet, use the panchromatic image for your Landsat scene, if it's Landsat 7? That should give you better results.

Thank you for the answer Whammer.