Antisense RNA (asRNA) strategies are identified as an effective and specific method for gene down-regulation at the post-transcriptional level. In this study, the major purpose is to find a correlation between the expression level and minimum free energy to enable the design of specific asRNA fragments. The thermodynamics of asRNA and mRNA hybridization were computed based on the fluorescent protein reporter genes. Three different fluorescent proteins (i) green fluorescent protein (GFP), (ii) cyan fluorescent protein (CFP) and (iii) yellow fluorescent protein (YFP) were used as reporters. Each fluorescent protein was cloned into the common pUC19 vector. The asRNA fragments were randomly amplified and the resulted antisense DNA fragments were inserted into the constructed plasmid under the control of an additional inducible plac promoter and terminator. The expression levels of fluorescent reporter protein were determined in real time by plate reader. Different results have been observed according to the fluorescent protein and the antisense fragment sequence. The CFP expression level was decreased by 50 to 78% compared to the control. However, with the GFP, the down-regulation did not exceed 30% for the different constructs used. For certain constructs, the effect was the opposite of expected and the expression level was increased. In addition, the YFP showed a weak signal compared to growth media, therefore the expression level was hard to be defined. Based on these results, a thermodynamic model to describe the relationship between the particular asRNA used and the observed expression level of the fluorescent reporter was developed. The minimum free energy and binding percentage of asRNA-mRNA complex were computed by NUPACK software. The expression level was drawn as a function of the minimum free energy. The results showed a weak correlation, but linear trends were observed for low energy values and low expression levels the CFP gene. The linear aspect is not verified for higher energy values. These findings suggest that the lower the energy is, the more stable is the complex asRNA-mRNA and therefore more reduction of the expression is obtained. Meanwhile, the non-linearity involves that there are other parameters to be investigated to improve the mathematical correlation. This model is expected to offer the chance to "fine-tune" asRNA effectiveness and subsequently modulate gene expression and redirect metabolic pathways toward the desired component. In addition, the investigation of the localization of antisense binding indicates that there are some regions that favors the hybridization and promote hence the down-regulation mechanisms.