Cover crops play a critical role in providing agroecological services such as improving soil health, reducing erosion and nitrogen loss, and suppressing weeds, which are closely tied to their performance such as accumulated biomass. This study evaluated the Active Canopy Sensor (ACS) -214, an active proximal sensing device equipped with its own light-emitting red and near-infrared spectral reflectance sensors, a time-of-flight laser, and an ultrasonic sensor, for estimating winter cover crop biomass across 13 U.S. states from 2020 to 2024. We assessed 11 species from three functional groups – grasses (n = 797), legumes (n = 264), and brassicas (n = 181) – using Random Forest (RF) models and four cross-validation strategies. The ACS-214 showed moderate to strong prediction accuracy for grasses (R² = 0.51 – 0.64) and legumes (R² = 0.44 – 0.76), though performance declined in leave-one-region-out analyses (R² = 0.06 – 0.46), indicating limited spatial generalizability. Brassica models had low prediction accuracy for all models (R² < 0.30), likely due to flowering and patchy growth. Biomass prediction breakpoints were observed at ∼3000 kg ha⁻¹ for legumes and ∼4000 kg ha⁻¹ for grasses. We also evaluated the effectiveness of using ACS-214 data to train Sentinel-2 satellite imagery for estimating grass cover crop biomass using withheld, out of bag data from 2023 to 2024. Sentinel-2 RF models trained with ACS-214 data showed good agreement with field-sampled (R² = 0.58 – 0.61) and ACS-214-estimated biomass (R² = 0.70). While Sentinel-2 offers scalability, the ACS-214 enables finer-resolution biomass mapping and better accounts for within-field variability, making it an effective tool for localized management and monitoring. These findings support the integration of proximal and satellite sensing approaches to enhance cover crop biomass estimation and agroecological assessment.