The predictive ability of partial least squares regression (PLSR) as calibration technique in fluorescence spectroscopy with regard to asphaltene and resin content was explored in some medium and heavy Venezuelan crude oils. These samples dissolved in toluene exhibited significant fluorescence signal at very low concentrations due to fluorophores associated with asphaltenes and resins. Additionally, crude oils spectra presented small shifts in the maximum peak position and small differences in the full width at half-maximum. This fact made the simple linear calibration technique meaningless for quantification purposes. Accordingly, multivariate calibration was applied to establish the relationship between the nonselective fluorescence spectra and the hydrocarbon content. Results confirmed that fluorescence spectroscopy with PLSR is a very promising method to predict the asphaltene and resin content of the crude oil in a fast and reliable way. Prediction results are comparable to those obtained by very cumbersome methods such as traditional SARA fractionation procedure.