Use UV spectrometry to determine under which wavelength the target compounds have the max absorption.

If you know the structure of the molecule, the molecule having conjugation or delocalization bonds in there structure shows UV light absorption.  If it is UV active you check the suitable   wavelength maxima using UV spectrophotometer.

The molecule doesn’t having UV activity, UV detector is not useful. Select other detectors like RI, PDA or ect

Use UV-Vis spectrophotometry to determine the wavelength at which the target compound shows maximum absorbance and also check the solvent effect on the absorbance maxima of the target compounds. It is helpful while choosing the mobile phase i. Acetonitrile, Methanol etc...

In addition, you can also use a Fluorescence Detector for compounds with low absorbance at UV-Vis. In this case, you can check literature in order to select a pair of wavelenghts (excitation and emission) suitable for the target compounds.

Take absorbance of the molecule of interest at different wavelengths and use the absorption maxima for your HPLC analysis

If you are using PDA detector. on selecting the peak of the drug, you will get the absorbance of the molecule, provided the compound is free of impurities, else, scan the molecule using UV spectrophotometer. Enter the wavelength in the detector option.

By using UV-Visible spectrophotometer you can find out the wavelength at which the compound of interest has maximum absorbance. If the molecule do not show UV activity, try with other detectors like PDA or RI or fluorescence detector, etc.

For more info, please refer to these links:

"Proper Wavelength Selection for HPLC Method Development (or Purity Determination)"; http://hplctips.blogspot.com/2013/08/wavelength-selection-for-method.html

"Method Development Hint: Use your HPLC Diode Array Detector (DAD or PDA) as a Spectrophotometer".

http://hplctips.blogspot.com/2013/09/method-development-hint-use-your-hplc.html

You asked how to select the wavelength for UV/VIS HPLC detection? To do so you must first scan the entire range of wavelengths of your sample (or sample mixture). Be sure to take into account any contributing absorbance of the solvent(s) that you dissolved the sample into (use a solution which has very low UV absorbance). Review the spectral data obtained and select the wavelengths which show clear "peaks" to start with. Next, run an analysis.

For HPLC we first separate and then analyze the sample peak(s) using a DAD (aka: PDA) to find the maximum peak wavelength in which we can reliably use to detect the sample (the apex). We use the full spectral data obtained for each peak to find this apex. Ideally, the wavelength selected will have little to no interference from other peak sample wavelengths and provide a strong signal with a narrow bandwidth. This is called the lambda max wavelength, "λmax". Each compound will have one or more.  The UV Spectra obtained for each peak can then be used to compare peaks with different or even similar retention times. The differences in their spectra can tell us something about their chemical bonds and structure.  Also, the spectral data is very useful for ruling out and troubleshooting the ID of peaks seen in the chromatogram (Note: Identical spectra does not equal identical samples as it is a null test of sorts. Different spectra = different samples and identical spectra = maybe the same/maybe not) ). For example: True enantiomers of the same racemate must have identical spectra or they are not enantiomers. Spectra of intermediates or reaction products can often be ID'd quickly by their spectra when following reaction chemistries by HPLC. Spectral data is very useful and why we must use a scanning DAD/PDA for all method development (for samples with chromophores). The DAD provides you with a tool to "see" many compounds during the development of the method for accuracy and also later on for QC and validation. Use of a single-wavelength detector can provide very misleading data as it will show a "peak" at a particular set wavelength, but that same peak will probably appear at many other wavelengths too.

It is important to evaluate the absorption spectrum of the molecule of interest. From it you should prefer a wavelength that is a apex or a valley in the spectrum. This minimizes the spectral interference generated by the measurement error of the detector itself. Choosing a apex in the spectrum will maximize the sensitivity of your method. But be careful not to exceed the linear range of the detector. When more than one substance is quantified in the same method you should also prioritize a wavelength at which molecules absorb proportionally, comparing the spectra of different molecules and looking for equivalent regions in the spectra. If you simply choose the maximum absorption of one substance and ignore the absorption of the others, your quantification will not be accurate.