It's well known the potential arrival of the AH1N1 flu disease can be realized in any large city, as well as its unknown impact and consequences on the people. Depending upon the strength of the propagation of virus, clearly it might fade away any scheme of preparedness has been designed. The experience of the 2009 worldwide flu pandemic have served to improve and test newest methodologies that target to toughen the resilience of the public health systems. In this paper we focus on the usage of a Machine Learning algorithm as an advantageous computational system aimed to support fast and effective decisions in epochs where a flu virus has initialized its spreading in a large or middle-size city. For this end the algorithm uses the formalism of the Wiener series that allows us to estimate predictions and thus manage decisions through these computational methodologies. In order to test the efficiency of the algorithm we used the 2009 Peruvian data where the flu A(H1N1) was spreading in Lima city with a velocity of 40 cases per week. We present simulations by which the usage of Machine Learning algorithms might be of importance to minimize undesired errors and optimize resources of public health services on those epochs where the velocity of spreading and number of contagious reaches their top values.