Are you using UV or fluorescence detection? The sample matrix or a constiuent of the sample has a lower absorbance than the mobile fase.

Try injections of different dilutions of the sample with the mobile phase to see the effect on the negative peak.

Hundreds of causes. More info needed. What is the method? Gradient or Isocratic? Mobile phase composition?  Injection solution and injection volume used (inj solutions which do not match mobile phase are common reasons for neg peak and/or poor peak shape/chromatography)? Improper use of the "Reference Wavelength" software feature (also, very common mistake). Which HPLC system? What is the actual HPLC method and settings? What are the retention time of the negative peak vs the sample / actual peaks? What is the peak amplitude? Can you include a chromatogram? What are the column dimensions and flow rate?

I think that you have a problem in setting of your instrument , you can get a negative of positive peaks according to what you choice 

Dear Sir,

Suggesting solution is very difficult based on the current information you have provided. Please provide following details .

1) Detector specs 

2) Mobile phase composition etc

However, if you are using UV detector presence of negative peak indicates analyte having less absorbance than mobile phase (background). Since it only comes in sample and not in the blank means it is present in sample only. Make sure that blank is having same composition as like sample except main analyte.  

thanks for your answers 

Please get the detail information for your ready refernce.

In HPLC, we are getting negative peak in sample solution, but not observed in blank (diluent) at 2.5 minute RT. What are the possible causes ? Detector is UV and mobile composition is 20mM Na2HPO4 pH 7 with OPA: Acetonitrile (95:05 v/v) at at flow rate of 1.0 ml/min in isocratic mode. Injection were made with 20 µl and main peak elute at RT 6.0 min.

Negative peak appeared because of the phase separation and new phase is non-signalling (Absorption, emission ,transmission etc..). a sample solution composition can be evaluate once with mobile phase or sample can be prepared once in mobile phase or some precipitation phenomenon may result this also within stystem... 

Am experiencing the similar problem. my mobile phase A(50mM acetate buffer pH 6.8) B (methanol) , gradient, injection volume 20 uL,

Gbolahan: Please start a new post. Do not reply here in someone else's. Include DETAILED information about your method and address the questions I posed in my response above.

William Letter posed some very good questions.

Very important are the following aspects:

1) Retention time of the negative peak, was there any retention by the HPLC conditions (column and mobile phase) applied?

2) Your mobile phase is phosphate buffer / acetonitril, hence showing some UV absorbance especially if you do not use the highest quality acetonitril.

3) What is the matrix of your sample, if it is aquous it is different from your mobile phase and that might be the cause of your negative peak.