Saliva is of interest as a diagnostic aid for oral and systemic diseases, to monitor therapeutic drugs, and detect illicit drug abuse. It is also attractive for biological monitoring of exposure to hazardous solvents. The major advantage of this indicator over other biological monitoring targets is that the saliva is noninvasive and less confidential in comparison with blood and urine. Salivary analysis is generally acceptable by study subjects and can be applied to investigation of a wide variety of compounds. However, very few studies have been conducted on the saliva matrix to monitor exposure to hazardous solvents. The aim of this study is to establish an analytical method, headspace solid-phase microextraction (HS-SPME) followed by gas chromatography–mass spectrometry (GC–MS), by which the saliva matrix can be monitored for multiple compounds with various polarities, such as methyl ethyl ketone (MEK), isopropyl alcohol (IPA), and N, N-dimethyl formamide (DMF) (common solvents used in synthetic leather manufacture), as well as acetone (ACE) and N-methyl formamide (NMF) (metabolites of IPA and DMF, respectively). We studied this technique as an alternative biological monitoring method for investigating exposure to hazardous solvents. A Carboxen/Polydimethylsiloxane (CAR/PDMS 75 μm) fiber coating was employed for this study, and various extraction and desorption parameters were evaluated. The extraction efficiency and reproducibility of analyses was improved by pre-incubation. The limits of detection were 0.004, 0.003, 0.006, 0.05, and 0.10 μg/mL for ACE, MEK, IPA, DMF, and NMF, respectively. Method validation was performed on standards spiked in blank saliva, and a correlation was made between HS-SPME and traditional solvent pretreatment methods. It was found that correlation coefficients ( r) were greater than 0.996 for each analyte, with no significant differences ( p > 0.05) between two methods. However, the SPME method achieved lower limits of detection, with good accuracy (recovery 95.3–109.2%) and precision (1.17–8.22% CV) for both intra- and inter-assay, when quality control samples were analyzed for all five compounds. The partition coefficient for each compound between the headspace of the saliva sample and the CAR/PDMS fiber coating was 90.9, 170.1, 36.4, 3.70 and 0.92 for ACE, MEK, IPA, DMF and NMF, respectively. Real sample analyses were performed on workers in a synthetic leather factory. In summary, the SPME method is a highly versatile and flexible technique for chemical measurement, and we demonstrate its application for monitoring biological exposure to hazardous solvents. Saliva monitoring using sensitive SPME approaches for determining workplace exposure should prove useful as an alternative exposure monitoring method.