<p>Residues of biodegradable plastics (BP) may be used for soil conservation techniques in deforested areas, but their effects on rill detachment—the most common erosion process on steep and long hillslopes—have never been explored so far. Furthermore, no equations have been proposed to predict the rill detachment capacity (<i>D</i>_<i>c</i>), rill erodibility (<i>K</i>_<i>r</i>), and critical shear stress (<i>τ</i>_<i>c</i>) in soils treated with these residues. To fill these gaps, this study evaluates <i>D</i>_<i>c</i> and its relationships with many physico-chemical and biochemical properties of soils treated with BP in comparison to untreated sites. Flume experiments have been carried out on soil samples collected in deforested sites of Northern Iran, and multivariate statistical techniques have been applied (principal component analysis, PCA, and agglomerative hierarchical cluster analysis, AHCA). Regression models have been developed to predict <i>D</i>_<i>c</i>, <i>K</i>_<i>r</i>, and <i>τ</i>_<i>c</i> for both treated and untreated soils. The BP application significantly increased the organic matter (+90%) and aggregate stability (+210%) and reduced the bulk density (−15%) of the treated plots compared to the untreated soils. Thanks to the effects on these key physical properties, <i>D</i>_<i>c</i> significantly decreased in the treated soils (by about 90%). PCA found close associations between <i>D</i>_<i>c</i> and soil properties, and AHCA clearly discriminated the BP-treated from untreated soils. The linear regressions between <i>D</i>_<i>c</i> and <i>τ</i> accurately estimated <i>K</i>_<i>r</i> and <i>τ</i>_<i>c</i> under treated and untreated soils. The most accurate predictor of <i>D</i>_<i>c</i> was the stream power under both conditions (treated and untreated soils), when a power function was used (Nash and Sutcliffe’s coefficient of efficiency &gt; 0.87).</p>