Background: The assumption of visceral leishmaniasis (VL) control program is that human disease (HVL) is related to canine infection (CanL), and that supports the culling of infected dogs. Objectives: Estimate the coverage of visceral leishmaniasis (VL) control measures and to relate them to the occurrence of HVL in an endemic urban area (Araçatuba, SP). Determine the CanL seroprevalence and to evaluate its relationship with the characteristics of dogs and their owners. Evaluated the culling dogs efficacy in controlling the canine infection. in the municipalities of Araçatuba and Birigui, state of São Paulo, Brazil. Methods: The cases of HVL and households with seropositive dogs registered in the period from 2007 to 2015 were geocoded the coverage of the serological inquiry, culling dogs and chemical spaying, canine seroprevalence and HVL incidence rates were calculated. The relationship between CanL, HVL and control measures was evaluated by map comparison, by linear regression and was also assessed using the Ripley K function. A cross-sectional study was conducted between 2015 and 2016, based on a sample of the canine population. Seroprevalence was modelled by logistic regression in a geostatistical approach using the integrated nested Laplace approximation for Bayesian inference. The spatial component was modelled by a Gaussian field, using a stochastic partial differential equation approach. Secondary data and the surveys were used to elaborate and calibrate dynamic models of canine disease transmission. Results: Verified the decrease of HVL and the CanL in Araçatuba over time, even in low coverage of control measures. The CanL seroprevalence was 8% in Araçatuba and 4% in Birigui. The occurrence of a seropositive dog in Araçatuba was associated with the presence of more than 10 dogs living in the same house, house with dogs that previously died of VL or died of other unnatural causes, and the place of dogs stayed during the day. Spatial dependence among observations occurred within about 46 m. Considering a continuous control and a serological survey effort three times higher than the average of the observed in Araçatuba and twice as high in Birigui, the activities related to culling dogs is effective in controlling canine infection. Conclusions: The observation of VL decrease over time may be related to control measures. The short-distance spatial dependence could be related to the vector characteristics, producing a local neighbourhood VL transmission pattern. The culling dogs were effective in controlling the disease in the dogs' population. However, this result should be understood with caution given the operational complexity of this measure and related ethical issues. New studies need to be developed for a better understanding if factors other than control activities would be involved in reducing the incidence of VL. It is necessary in addition to long-term planning of the control actions carried out by the program, investment in researches that evaluate the cost-effectiveness of other measures that may...