In this paper, we extend the differentiated deployment method which we have proposed in our previous paper by adding the connectivity constraint. We assume a fixed communication ray and a probabilistic detection model. We also consider that to each point in the deployment area is associated a detection probability threshold, which must be satisfied by our deployment method. Finally, we suppose that the detection probabilities thresholds of the area are geographically nonuniformly distributed. Our differentiated deployment problem is modeled as multi-objectives optimization problem, which we resolve using our proposed Tabu Search-based algorithm. A comparison is made with the methods found in the literature. The performances obtained by our method are much better in term of required number of sensors, generated detection probabilities, while ensuring the network connectivity.