**Background:** Adolescence is a critical period for establishing healthy eating habits that prevent diet-related diseases later in life. Worldwide, the majority of adolescents consume poor-quality diets with insufficient fruits, vegetables, and calcium-rich foods, and excessive fatty foods, sweets, and soft drinks. In Ethiopia, adolescents constitute about one-fourth of the population, yet data on their diet quality are limited. Good-quality diet during adolescence is essential for optimal health, physical growth, intellectual development, academic performance, and prevention of non-communicable diseases.
**Methods:** A school-based cross-sectional study was conducted from 1 February to 20 February 2019 in Awi Zone, Amhara Region, Ethiopia, which had 16 preparatory schools with 15,250 students. A sample of 846 students was calculated using a single population proportion formula (95% confidence level, 5% margin of error, 50% assumed proportion of good-quality diet, design effect of 2, and 10% non-response rate). Multistage sampling was used: 4 schools were randomly selected from 16, then a population-proportionate sample was drawn, and students were selected via simple random sampling from school registers. Data were collected using a pretested, self-administered structured questionnaire. A 40-item food frequency questionnaire assessed dietary intake over the previous 3 months; students who consumed foodstuffs at least once weekly were considered consumers. Diet quality was a composite measure: students with high dietary diversity score (highest tertile of food groups consumed), appropriate animal-source food consumption (highest tertile), and adequate food variety score (above mean) were classified as having good-quality diets. Bivariate and multivariable logistic regression analyses were performed using SPSS V.25.0; variables with p≤0.2 in bivariate analysis were entered into the multivariable model, and statistical significance was set at p<0.05.
**Key Results:** Of 846 invited students, 834 participated (response rate 98.6%). Mean age was 18.8 years (SD ±1.06); 52.9% were male; 73.7% were from rural areas; 78.9% of mothers and 74.1% of fathers had no schooling; 68.6% of households had monthly income ≤4000 Ethiopian birrs. Only 24.7% (95% CI: 21.7%, 27.7%) of students had good-quality diets. Dietary diversity was achieved by 34.9%; adequate food variety score by 39.8%; and appropriate animal-source food consumption by 30.2%. Regarding meal frequency, 72.1% had three or more meals daily. Consumption of unhealthy items included biscuits (22.3%), doughnuts (15.6%), and soft drinks (16.2%). All students consumed cereals daily; nearly three-fourths consumed legumes daily; but animal-source foods, fruits, and vegetables were seldom consumed. Factors significantly associated with good-quality diet in multivariable analysis: female sex (AOR=2.88, 95% CI: 2.0, 4.1), urban residence (AOR=1.90, 95% CI: 1.1, 3.2), mother having schooling (AOR=1.78, 95% CI: 1.1, 2.7), having pocket money (AOR=1.83, 95% CI: 1.2, 2.6), receiving nutrition information (AOR=1.90, 95% CI: 1.2, 3.1), household monthly income 4001–8000 birrs (AOR=2.85, 95% CI: 1.8, 4.5), and household monthly income >8000 birrs (AOR=3.90, 95% CI: 2.2, 7.1). The Hosmer-Lemeshow goodness-of-fit test result was 0.757.
**Clinical Implications:** The vast majority (75.3%) of preparatory school students in this Ethiopian zone consume poor-quality diets, with particularly low consumption of animal-source foods, fruits, and vegetables. This places adolescents at risk for micronutrient deficiencies, impaired growth, poor academic performance, and future non-communicable diseases. The strong association between household income and diet quality (nearly 4-fold higher odds for the highest income bracket) underscores economic barriers to healthy eating. Nutrition education that accounts for sociodemographic characteristics, along with income-generating interventions for low-income households, is urgently needed. The study is limited by its cross-sectional design (cannot establish causality), use of self-reported dietary data (potential social desirability bias), and lack of generalizability to out-of-school adolescents.