Heart disease is one of the leading causes of death worldwide. Among the various methods used to assess heart function, the MPI SPECT method is a valuable and non-invasive method that brings high-quality images with low radiation exposure. Radiomics has been developed to extract quantitative features from medical images. These features can be used to predict diagnosis and treatment in medical science. To use these features in the clinic, they need to be reliable; in other words, they need to be repeatable and reproducible. Various factors, including different reconstructions, can affect the repeatability and reproducibility of radiomic features. Twenty patients who underwent stress and the rest of SPECT MPI were used in this study. As a result, 40 existing images were reconstructed in 15 different modes. Finally, 600 unique reconstructions were obtained, and the segmentation process was conducted using the 3D-Slicer program. Feature extraction was done using LIFEx, and finally, the coefficient of variance (COV) method was used to check the reproducibility. The most robust features were FO_Kurtosis, GLCM_Entropy_log10, GLCM_Entropy_log2, GLRLM_SRE, GLRLM_LRE, GLRLM_RP, GLZLM_SZE, and GLZLM_HGZE. The change in the order reconstruction parameter was the only case that caused the least feature variation. This study is planned to assess the reliability of radiomic features from MPI SPECT images over changes in reconstruction parameters.