Abstract: Current research has been demonstrating the roles of miRNAs in a variety of cellular mechanisms, lending credence to the association between miRNA dysregulation and multiple diseases. miRNA molecules function through a complementary binding with messenger RNAs (mRNAs), induced post-transcriptional repression. Several medical applications, including miRNA mimics and miRNA inhibitors, have been proposed to treat diseases by interfering with miRNA functions. Understanding the mechanisms of miRNA is critical for developing effective diagnostic and therapeutic strategies. Due to experimental constraints, several computational-based models have been developed over the years to facilitate the prediction of miRNA-mRNA interactions, albeit they present limited predictive capabilities, poor characterisation of miRNA-mRNA interactions, and low accessibility and scalability. To address their drawbacks, we developed a PRedictive model for Identification of novel MIRNA-Target mRNA Interactions (PRIMITI), a novel machine learning model that utilises CLIP-seq and expression data to characterise functional target sites in 3’-untranslated regions (3’-UTRs) and predict miRNA-target mRNA repression activity. Negative sample selection, an approach to provide more reliable negative samples, was also implemented, coupled with a newly introduced SNP information and iFeatures. PRIMITI achieved correlation coefficients up to 0.80 for a prediction of functional miRNA-target site binding and up to 0.76 for a prediction of miRNA-target mRNA repression activity on cross-validation and an independent blind test. Additionally, the model succeeded in predicting repression in an unseen microarray dataset and collection of validated miRNA-mRNA interactions, yielding the largest number of validated miRNA-target repressions accurately predicted as repressed, when compared to state-of-the-art methods. PRIMITI was made publicly available on a reliable, scalable and user-friendly web server at https://biosig.lab.uq.edu.au/primiti/.