To describe the feasibility, acceptability and results of Strong Families Start at Home, a 6-month pilot trial of a home-based food parenting/nutrition intervention. Pilot randomised controlled trial. Participants received six visits with a community health worker trained in motivational interviewing (three home visits, three phone calls); an in-home cooking or reading activity; personalised feedback on a recorded family meal or reading activity; text messages and tailored printed materials. Parents and their 2-5-year-old child were randomised into intervention (responsive food parenting practices/nutrition) or control (reading readiness) groups. Parents (n 63) were mostly mothers (90 %), Hispanic/Latinx (87 %), born outside the USA (62 %), with household incomes <$25 k (54 %). Despite delivery during COVID-19, 63 % of dyads were retained at 6 months. The intervention was delivered with high fidelity. All parents in the intervention group (n 24) expressed high levels of satisfaction with the intervention, which produced positive treatment effects for whole and total fruit component Healthy Eating Index-2015 scores (point estimate (PE) = 2·14, 95 % CI (0·17, 1·48); PE = 1·71, 95 % CI (0·16, 1·47), respectively) and negative treatment effects for sodium (PE = -2·09, 95 % CI (-1·35, -0·04)). Positive treatment effects also resulted for the following food parenting practices: regular timing of meals and snacks (PE = 1·08, 95 % CI (0·61, 2·00)), reducing distractions during mealtimes (PE = -0·79, 95 % CI (-1·52, -0·19)), using food as a reward (PE = -0·54, 95 % CI (-1·35, -0·04)) and providing a supportive meal environment (PE = 0·73, 95 % CI (0·18, 1·51)). Given the continued disparities in diet quality among low-income and diverse families, continued efforts to improve child diet quality in fully powered intervention trials are needed.