Abstract Purpose: Recombinant proteins have become increasingly important items in research and industry. Due to its low cost, high yield and rapid growth rate, Escherichia coli (E. coli) is the first choice as host for the production of recombinant proteins. The expression of recombinant proteins in E. coli systems often result in inclusion bodies lacking proper folding and structure. In silico bioinformatics prediction tools may be promising in optimal expression of soluble recombinant proteins. Materials and methods: In this review, we aimed at making critical recommendations on how to improve the soluble expression of recombinant proteins. Furthermore, we compared the solubility of recombinant proteins using bioinformatics prediction tools versus experimental results. Data were analyzed using SPSS software. Results: Some recommendations worthy of consideration in gene design and expression were reminded. The results of a comparison between bioinformatics and experimental methods revealed that no significant coordination existed. RPSP and SOLpro showed higher sensitivity (43.5% and 56.5%, respectively) and specificity (52.9% and 47.1%, respectively), when compared to FoldIndex and PSoL. The results from p-value and roc curve indicated the effect of MW, helix percentage and aliphatic index on protein solubility (p-value< 0.05). Conclusions: This review discusses efficient expression of soluble recombinant proteins. The bioinformatics prediction tools were examined for their sensitivity and specificity. MW, helix percentage and aliphatic parameters should be considered in gene design.