The rapid increase in the use of pesticides is driven by the growing demand in the agricultural sector. However, the widespread application of these pesticides and their inherent toxicity have significant repercussions on the ecosystem, particularly impacting animal and bird species. In this present study, we have developed four 2D quantitative structure-toxicity relationships (QSTRs) models for four different avian species using the largest number of available experimental data points to date employing the partial least squares (PLS) algorithm. Furthermore, we have also performed the Read-across algorithm to improve the test set results. Based on the information derived from the models, it was found that hydrophilic characteristics, the presence of molecular branching and thio imide groups impact negatively pesticide toxicity, while the presence of phosphate group, presence of halogens viz. chlorine and bromine atoms, presence of hetero atoms, high molecular weight, presence of bridgehead atoms, presence of secondary aliphatic amide and fragments like NCOO escalates avian toxicity. The developed QSTR models were further employed to predict the Pesticide Properties DataBase (PPDB) for all four avian species as a measure of data gap-filling and risk assessment. Thus, the developed models can be utilized for eco-toxicological data-gap filling, prediction of toxicity of untested pesticides as well as the development of novel and safe environmental-friendly pesticides.