The validity of the field in situ spectral model under the influence of many environmental factors in the field has become a key scientific problem that needs to be broken through in order to promote the technology on a large scale. In this work, 158 soil samples were collected from the Huangshui River Basin of the Qinghai-Tibet Plateau, Then the spectra and Organic Matter of these soil samples were measured under in-situ field and laboratory conditions, re-spectivel. Based on the in-situ field spectra processed by direct standardization (DS) algorithms, piecewise direct standardization (PDS) algorithms, general least squares weighting (GLSW) and orthogonal signal correction (OSC) algorithms, the laboratory spectra, field spectra, field DS spectra, field PDS spectra, field GLSW spectra and field OSC spectra were used to estimate the soil organic matter content by stepwise regression models of principal components, and the stability of each model was cross-validated. In order to further prove the effectiveness of the model with combined laboratory and field spectra, the performance of this model was compared with five types of hyperspectral remote sensing models including those with combined laboratory and in-situ field, combined laboratory and DS transformed field, combined laboratory and PDS transformed field, combined laboratory and GLSW transformed field, combined laboratory and OSC transformed field, one by one. The results showed that the precision of the Field-PCSR model (R² = 0.53, RPD = 1.17) was significantly lower than that the one with laboratory spectra(R² = 0.90, RPD = 2.26), and the precision of the model with DS, PDS, and OSC transformed in-situ field spectra was improved (RPDDS=2.23; RPDPDS=1.62 ; RPDosc=1.32). The precision of field DS, field PDS, field GLSW and field OSC spectra combined with laboratory spectra improved significantly, the model with combined laboratory and DS transformed in-situ field spectra is the one with highest accuracy, the RPD is up to 2.26. The results confirm the rule of the priori knowledge of laboratory spectra and DS, PDS, GLSW, and OSC algorithms can improve the reliability of models for oil organic matter content estimation, which can provide important theoretical and methodological support for the development of in-situ field hyperspectral remote sensing detection of soil organic matter content. [ABSTRACT FROM AUTHOR]