Purpose: In Cambodia, Village Animal Health Workers (VAHW) have been trained by NGOs or by the government to provide animal health services (treatment, husbandry advice, vaccination) to their communities' farmers. This system is characterised by a high variability of skills because of non-harmonised training plans, poor sustainability with a large number of VAHWs dropping their activities after five years and no harmonised tools for their evaluation. The objective of the study was to assess the work skills of VAHW. Methods: We applied a scoring grid composed of five categories (sustainability, treatment, production, vaccination, reporting). Linked to several evaluation criteria, specific questions were defined to assess if the criteria were fulfilled by the VAHW. An additional questionnaire, with 31 explanatory variables, was developed in order to collect data about factors that could influence the VAHW's score. The study was implemented in three provinces bordering Vietnam (Kampong Cham, Prey Veng and Takeo). A total of 367 villages were selected using a proportional random sampling, g method. We applied a multivariable linear regression model to determine factors associated with high scores for the VAHW evaluation. Results: In the population studied, 23% of the villages did not have a VAHW. According to our scoring system, 23.6% of the VAHW interviewed were in a situation of inactivity. From our multivariable analysis, six factors were significantly associated with a high score in the evaluation of the VAHW once they were active: selecting a VAHW from a village with at least 100 heads of cattle, using practical activities during training, having a training duration longer than 30 days, organising refresher courses, being a member of association and having regular contact with the district veterinarian. Conclusions: These results demonstrate the need of constant networking activities in the surveillance system to ensure that field staffs do not feel isolated. Relev