Armstrong’s view on a law of nature stands between regularity theory and metaphysical necessity theory. He maintains that a law of nature is necessary on the one hand, and contingent on the other hand. He holds that a law of nature is necessary in the sense that it is beyond regularity. It is meant by ‘contingent’ that it does not hold in all possible worlds. A law of nature in our world may not hold in other possible worlds. In other words, he is defending that a law of nature has a physical necessity. The necessity of a law of nature is said to be required, since there are at least three main difficulties. The view on physical necessity is maintained with Armstrong’s categoricalism on the nature of properties. According to it, properties do not have modality as such, which is externally attributed by a law of nature. I argue, in this paper, that there are problems in his defending of the physical necessity of a law of nature in his system, which ranges over his categoricalism, quidditism, actualism and combinatorialism. My criticism on physical necessitarianism is different from others’ in that it considers the whole system of Armstrong’s views on relevant topics. In the first part, I present what physical necessity means in Armstrong’s theory of laws. In the second part, I point out two main problems of his physical necessitarianism in considering his views on the relevant topic such as the nature of properties, the criterion of existence and possible worlds. In 2.1,I reformulate his arguments and argue that they fail mainly because of his categoricalism. He tries to defend the physical necessity of laws by means of universals. I show that his arguments are circular. In 2.2, I show that he has problems in admitting the physical necessity as primitive. I argue that the view on the physical necessity is not coherent with his views on possible worlds, as well as those on the criterion of existence and the nature of properties.