The way to normalize them is to calibrate them to radiance or TOA reflectance, but preferably an atmospheric correction to surface reflectance. Then the data will be in the same 'units' of reflectance and related to the actual physical properties of the surface. If you are only trying to make them visually appealing, then you can use Stretch Data to convert them to the same range of your choice.

Thank you for prompt input.

Actually, I am interested to perform 'Relative Radiometric Correction' in spite of 'Absolute Correction' which often involves atmospheric correction to produce normalized multi-temporal datasets. While, I am intending to perform 'Scatter Plot Matching (Maas and Rajan 2010' technique for which it is necessary to normalize digital counts prior to compare Red and NIR bands of two dates, So, as a preliminary step, I have to normalize OLI with ETM+ which has lower radiometric resolution (0-255).

As an example, I have reported existing Min and Max values of two bands from different sensors. How would I transform 'Min' values of OLI prior to populate 'Output Min' field of the 'Stretch' tool to derive the resultant data stretch keeping in view existing Min value i.e. 8730 (OLI band) ? Surely, Max field will be populated with the value of 255.

If you want to 'match' the ETM data that sounds like it is byte-scaled from 0-255, then in the Stretch Data tool, set the output data type to byte and set the min/max to 0/255. You can even stretch the ETM data to the full range of 0-255 to make sure they both are scaled linearly within this range.

Thank you. It sounds nice to re-scale ETM+ Min/Max to 0/255 prior to match OLI stretched bands on the same scale.