Objective To analyze the possible factors influencing serum albumin in the elderly. Methods A total of 267 elderly individuals (≥60 years) without serious disease were offered a clinical evaluation with physical examination. General data, such as weight, past medical history, residential area(urban or rural), vocation, diet, daily intake, marriage, sleep and mental state were recorded. Meanwhile, serum albumin, total protein, alanine aminotransferase, and serum creatinine were assessed. The subjects were grouped into categories according to different factors. Results Serum albumin was significantly lower in the oldest old group than in normal elderly group[(35.6±3.4) vs (38.2±3.2)g/L, P＜0.01], in male group than in female group[(36.4±3.5 ) vs (38.2±3.2)g/L, P＜0.01], in vegetarian diet group than in meat-based diet group[(34.8±2.1) vs (37.6±3.5)g/L, P＜0.01], in low intake group than in normal intake group[(35.4±3.6) vs (37.7±3.3)g/L, P＜0.01], in individuals who live alone than in those who live together with a partner[(35.6±3.4 )g/L vs (37.9±3.3 )g/L, P＜0.01], in insomnia group than in good sleep group[(36.0±3.4) vs (38.2±3.2)g/L, P＜0.01], in anxiety or depression group than in good mental state group[(35.7±3.8) vs (37.4±3.4)g/L, P＜0.01]. No significant difference was found between urban group and rural group, as well as between mental workers and physical workers. Linear regression analysis showed that serum albumin was negatively correlated with age(r =－0.196, P＜0.01), positively correlated with weight(r = 0.133, P＜0.01) and total protein (r =0.282, P＜0.01), but not correlated with alanine aminotransferase and serum creatinine. Conclusion Many factors, including sex, age, diet, daily intake, marriage state, sleep and mental state, may influence serum albumin in the elderly, which decreases with age and changes in parallel with weight and total protein.