Ensuring quality in pervious concrete poses challenges, limiting its use. This work investigates the potential of machine learning to forecast its properties, offering a novel and accessible approach. Five machine learning techniques were employed on 300 experimental data points, considering mix parameters (aggregate size, ratio, compaction) and non-destructive measurement (ultrasonic velocity, resistivity). Artificial Neural Networks (ANNs) excelled, achieving high accuracy (R2 > 0.97) for prediction of porosity and compressive strength. Sensitivity analysis revealed the dominant influence of compaction energy, aggregate-to-cement ratio, and ultrasonic velocity, while aggregate size and resistivity had minimal impact. This study suggests that machine learning models, particularly ANNs, can be reliable and efficient for predicting pervious concrete properties. This has the potential to improve quality control and encourage broader adoption in the construction sector, ultimately leading to more sustainable and permeable infrastructure.