Texture, and particularly firmness, is an important and commonly used indicator of internal quality and ripeness of fruit. An objective method of measuring firmness would be beneficial in achieving quality control and reducing costs in fruit marketing, especially for export. Reflectance near infrared hyperspectral imaging (NIR-HSI) has been used as a nondestructive method for this purpose and was tested on predicting the firmness of pineapples. The spectral data, using NIR-HSI, and firmness, using a texture analyzer, were tested on 120 pineapples in order to develop a calibration model using partial least squares regression (PLSR). Samples were divided into a calibration set (N = 80) and a prediction set (N = 40) for establishing and testing the model, using the best method from spectral pretreatments was by multiplicative scatter correction (MSC). The accuracy of the model had Rp and RMSEP of 0.77 and 0.15 kgf, respectively indicating that there was good correlation between NIR-HSI and texture analyzer measurement showing that the model could be used for screening intact pineapples for firmness nondestructively.